{"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nJapan had surrendered .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"surrendered\", \"conflict.yield.n/a\", \"An entity ends active conflict by yielding, indicating a cessation of hostilities through capitulation or submission.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe commodities trader who “ bought ” his appointment to a Top Secret nuclear weapons security advisory board .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"bought\", \"transaction.transferownership.n/a\", \"Represents a transfer of ownership or control where one party acquires rights to an asset or position through purchase or exchange.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nKlong Prem is home to 6,267 inmates serving sentences that range from 15 years to life , with 64 percent convicted of drug - related crimes .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"convicted\", \"justice.judicialconsequences.n/a\", \"This event indicates a legal determination of guilt based on evidence and judicial process, reflecting a formal finding of criminal responsibility.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nI remember another trip to New York where we walked by the Trump Tower .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"walked\", \"movement.transportperson.selfmotion\", \"Describes a self-directed movement of a person through space, indicating physical travel or walking without external transportation.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMagliocchetti pleaded guilty in 2010 to involvement in a pay - for - play campaign finance scheme .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"guilty\", \"justice.judicialconsequences.n/a\", \"This event indicates a legal admission of responsibility for illegal activities involving campaign finance corruption.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe higher the percentage of Muslims in a society , the lower the homicide rate .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"homicide\", \"life.die.deathcausedbyviolentevents\", \"This event represents a death resulting from violent acts, specifically homicides, in a societal context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ We were sacked because the new government needs scientists who will applaud increased logging , to convince public opinion that this insane idea is okay , ” said Przemysław Chylarecki , one of the dismissed scientists .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"sacked\", \"personnel.endposition.n/a\", \"An employee is removed from their position due to organizational or policy-related reasons, often involving disagreement with management or ideology.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIt appears that none of the anti - Muslim critics are aware that Muslim societies are among the least violent in the world .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"violent\", \"life.injure.injurycausedbyviolentevents\", \"This event represents injury resulting from violent actions or events involving physical harm or conflict.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWhat Clinton said was false and misleading .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"misleading\", \"contact.prevarication.broadcast\", \"Broadcasting false or misleading information intentionally to mislead the public or influence opinion.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn Saddam Hussein 's Iraq that might work when opponents can be thrown in jail or exterminated .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"exterminated\", \"life.die.deathcausedbyviolentevents\", \"Death resulting from violent actions such as execution or mass killing in a political or military context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nEarlier in the campaign , Trump denigrated McCain 's status as a war hero , saying he prefers people who do not get captured .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"captured\", \"movement.transportperson.preventexit\", \"Prevents a person from exiting a location during transport, restricting their movement or departure.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nInformal meeting of EU 27 next week without PM in the room to decide common negotiating position vs UK on exit negotiations — Faisal Islam ( @faisalislam ) June 24 , 2016\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"negotiating\", \"contact.negotiate.n/a\", \"EU members meet informally to establish a unified stance on Brexit negotiations without the UK Prime Minister present.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nFor example , Democracy Now reported after the DNC Leaks ( resulting in the resignation of Debbie Wasserman Schultz and other DNC officials ) that “ WikiLeaks has not revealed the source of the leaked emails , although in June a hacker using the name Guccifer 2.0 claimed responsibility for the hacking into the DNC ’s computer network . ” Someone hacks into a computer network , or acquires the information , then gives this information to WikiLeaks .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"acquires\", \"transaction.transferownership.n/a\", \"An entity acquires control or information from a system, transferring ownership or access rights to another party.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nShe said she had met with Manafort twice but said neither Manafort nor Gates played a role in its lobbying activities .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"met\", \"contact.discussion.meet\", \"An event where two individuals meet to discuss matters, indicating interaction and potential exchange of information or opinions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMany of those who survived the initial blast on August 6 died of severe radiation - related injuries and illnesses .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"blast\", \"conflict.attack.bombing\", \"A bombing attack during a conflict causing physical destruction and casualties through explosive force and subsequent radiation effects.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAnd of course , an army of government shills and trolls have thoroughly infiltrated if not saturated both mainstream as well as alternative internet and social media with the expressed purpose of muddying the waters to intentionally misinform , deceive and confuse the public , making them highly susceptible to nefarious manipulation by fear porn propaganda .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceive\", \"contact.prevarication.correspondence\", \"An event where false information is deliberately shared to mislead, manipulate, and distort public perception through communication channels.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHomicides in Chicago are up 49 percent , according to Chicago Police Department figures , with 441 homicides as of Monday .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"homicides\", \"life.die.deathcausedbyviolentevents\", \"This event represents a person's death due to intentional violent acts, specifically homicide, involving a fatal attack by another individual.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nNo one would have ever thought voting would potentially move us closer to annihilating humankind , and it took Kevin Costner to pull us back from what seemed like certain doom .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"annihilating\", \"conflict.attack.n/a\", \"An event where a hostile action threatens total destruction of humanity, indicating a severe conflict or existential risk.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAnd this is based on the fact that there was a study by Pew which found that there were up to 1.8 million deceased voters on voting rolls .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceased\", \"life.die.nonviolentdeath\", \"An individual died without violence, indicating natural or non-aggressive causes of death in a demographic or electoral context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nDante Scala , an associate professor of political science at the University of New Hampshire , said Sen. Kelly Ayotte Kelly AyotteKelly Ayotte joins defense contractor 's board of directors Week ahead : Comey firing dominates Washington GOP senators pitch Merrick Garland for FBI director MORE ( R - N.H. ) , who faces a tough challenge from Democratic Gov.\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"firing\", \"personnel.endposition.n/a\", \"Denotes the termination of a person's position or role within an organization or institution.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn 2013 , Syrian hackers backing embattled President Bashar Assad hijacked The Associated Press ’ Twitter account , tweeting out falsified reports of two explosions at the White House that had injured Obama .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"hijacked\", \"conflict.attack.n/a\", \"An unauthorized takeover of a digital platform to spread false information during a political conflict.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMr Roslovtsev , 36 , has been detained as least nine times by Russian authorities after walking in Moscow 's Red Square in his Putin mask .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"detained\", \"movement.transportperson.preventexit\", \"An individual is prevented from exiting a transport or area by authorities, restricting movement due to legal or security reasons.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMORE said it ’s possible he could ask his former rival for the nomination , Sen. Marco Rubio Marco RubioOvernight Finance : Ryan seeks manufacturing muscle for tax reform | Warren targets Wells Fargo board | Senators raise concerns over Russian takeover of Citgo | Pelosi hits GOP for budget delays Senators ask administration to probe possible Russian takeover of Citgo Week ahead : Obama DHS chief to testify on Russian meddling MORE ( R - Fla . ) , to be his running mate .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"testify\", \"justice.initiatejudicialprocess.n/a\", \"This event type indicates the initiation of a judicial process, such as a legal proceeding or investigation, by a governmental or legal authority.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe primary goal , however , is to provide aid and protection to refugees in the countries and areas where they have fled , “ so refugees can return home when the conflict ends , ” the official added .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"provide aid\", \"transaction.transfermoney.giftgrantprovideaid\", \"This event represents the provision of financial or material support to refugees for their safety and eventual return home after conflict ends.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nYuri Y. Chaika , Russia ’s prosecutor general , publicly urged law enforcement agencies to be less lethargic .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"urged\", \"contact.requestadvise.broadcast\", \"An official calls on law enforcement to improve responsiveness and take proactive action in legal enforcement.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAccording to news reports from Alabama , eleven of Tahir Javed ’s convenience stores affiliated with criminal activity were confiscated over the summer of 2015 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"confiscated\", \"transaction.transaction.n/a\", \"This event denotes the seizure of business assets linked to illegal activities by authorities during a specific time period.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nPresident Obama said that in the wake of attacks in places including Brussels , there is “ not only great urgency around the nuclear issue but eliminating generally the scourge of terrorism . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"attacks\", \"life.injure.injurycausedbyviolentevents\", \"Violent attacks cause physical or psychological harm to individuals or groups in specific locations during certain times.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe clip comes from a 2010 interview on Political Vindication Radio , a conservative talk show in California .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"interview\", \"contact.discussion.meet\", \"An event where two parties engage in a discussion or conversation, typically involving exchange of ideas or opinions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nOf course , no one really believed that Naryshkin had read Inozemtsev ’s article or that he was guilty of copying it himself .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"guilty\", \"justice.judicialconsequences.convict\", \"This event signifies a legal finding that an individual is proven guilty of a crime or wrongdoing through judicial process.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWhile there is no record of a coalition air strike on Abu Kamal on the day the hospital was hit - 6 March - a coalition press release does talk about an oil pipeline nearby being bombed .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"strike\", \"conflict.attack.bombing\", \"A military bombing action carried out by a coalition force against a target in a conflict zone involving armed violence and destruction.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIt ’s no wonder that despite waving Islamist red flags everywhere from his place of employment to the FBI , Mateen roamed freely until he committed mass murder .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"roamed\", \"movement.transportperson.selfmotion\", \"Describes an individual's unmonitored movement through spaces without external transport or guidance, indicating free mobility.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nA group of prominent intellectuals and activists is defying France ’s crackdown on the Palestine solidarity movement by publicly calling for the boycott of Israeli goods .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"boycott\", \"transaction.transaction.embargosanction\", \"An official or organized refusal to engage in trade with a country or entity as a form of political or economic pressure.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nInmates sit on the floor during an inspection visit in the long term sentence zone inside Klong Prem high - security prison in Bangkok , Thailand July 12 , 2016 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"inspection\", \"inspection.sensoryobserve.physicalinvestigateinspect\", \"An official checks physical conditions of prison cells during a routine visit to ensure safety and compliance with regulations.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nGreat hate and sickness , \" the tweet read .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"sickness\", \"life.injure.n/a\", \"This event denotes a physical or mental harm caused to an individual, resulting in illness or injury, often due to external factors or conditions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe most recent Islamic act of war , in which self - professed Islamic State jihadi Omar Mir Seddique Mateen slaughtered 49 people and injured scores of others at a gay nightclub in Orlando , Fla. , should be a game changer .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"slaughtered\", \"life.die.deathcausedbyviolentevents\", \"This event captures deaths resulting from violent acts, indicating loss of life due to warfare or terrorism in a specific location and time.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn the coming presidential election , the outcome will probably be determined by whether the powerful oligarchic interest groups decide whether Trump is an actual threat or whether they can cosy up to him and rope him in by appointing his government .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"appointing\", \"personnel.startposition.hiring\", \"This event signifies the appointment of a person to a new position within an organization or government structure.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n\" While it would be our honor to offer the opportunity to observe our voting process , it is prohibited under state law to allow anyone except election officials and voters in or around the area where the voting takes place , \" Benge wrote .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"voting\", \"government.vote.n/a\", \"This event denotes a governmental voting process observation opportunity restricted by state law to specific authorized individuals only.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nPhotos : World War II in pictures A jubilant American sailor kisses a nurse in New York 's Times Square on August 14 , 1945 , as he celebrates the news that Japan has surrendered .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"surrendered\", \"conflict.yield.surrender\", \"A conflict ends when one side formally gives up resistance, marking the conclusion of hostilities and the cessation of military actions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe argument will include specific criticism of comments Mr. Trump has made about rethinking the United States ’s support of NATO ; his proposal to allow Japan , South Korea and Saudi Arabia to acquire nuclear weapons ; his vow to temporarily bar Muslims from entering the United States ; and his pledge to advance the use of torture and kill the families of suspected terrorists .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"vow\", \"contact.commitmentpromiseexpressintent.meet\", \"An event where a person publicly commits to a specific intention or action through a formal pledge or declaration.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nalleging that participants in the movement are violent , spiteful , unrealistic , self - defeating , ignorant , destructive / anarchistic , or in some way grievously biased ;\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"violent\", \"life.injure.n/a\", \"This event signifies a physical or psychological harm caused by violent actions, indicating potential danger or damage to individuals involved in the movement.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nShahran Amiri , the nuclear scientist executed in Iran last week , claimed to have been abducted to Arizona by the CIA in a YouTube video uploaded in June 2010 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"executed\", \"justice.judicialconsequences.n/a\", \"This event denotes a judicial or state-sanctioned execution, indicating a formal end to a person's life by authority.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe IAAF have voted to suspend Russia 's athletics team by 22 votes to one\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"voted\", \"government.vote.castvote\", \"A governmental body formally cast a vote to make a decision on a policy or action involving a nation's athletic team.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe importance of what he did for the country outweighs the law that he violated and the just move is a pardon for Edward Snowden .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"violated\", \"government.agreements.violateagreement\", \"An entity breaches a formal agreement or legal commitment made by a government or institution, indicating a violation of established rules or obligations.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThey 'll mutilate me and my uniform .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"mutilate\", \"life.injure.n/a\", \"This event represents a severe physical harm inflicted on an individual, involving damage to body parts or clothing, indicating a violent act against a person's physical integrity.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nGuccifer 2.0 had always claimed to be Romanian like the original Guccifer , imprisoned hacker Marcel Lazăr Lehel , but the 2.0 version had regularly communicated with journalists in Russian .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"imprisoned\", \"movement.transportperson.preventexit\", \"An event where a person is prevented from exiting a location, typically implying restriction or confinement.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nEventually the police were able to subdue and arrest Saidov 's attacker .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"arrest\", \"movement.transportperson.preventexit\", \"Preventing an individual from exiting a location through force or control during movement.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWhile the United States exports little food to Russia , some European Union countries had a significant trade relationship with Russia prior to the embargo .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"embargo\", \"transaction.transferownership.embargosanction\", \"An embargo restricts trade or transfers of goods between countries, often imposed due to political or economic reasons.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nFBI officials expect they have to interview Abedin again after they have gone through the emails .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"interview\", \"contact.discussion.meet\", \"An event where FBI officials plan to meet with Abedin for a discussion or interview regarding email content and possible investigations.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAmericans see the United States as a super - power but it is living on borrowed glory .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"borrowed\", \"transaction.transferownership.n/a\", \"This event represents a transfer of ownership without formal documentation or legal basis, implying a temporary or unsustainable arrangement.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe ISIL leader is indicated in the documents as the person behind the transportation of bombs from Syria to Turkey .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"bombs\", \"conflict.attack.n/a\", \"Represents an attack event without a specific defined nature or classification in the conflict context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nSome onlookers have suggested that the alleged thieves could be linked to Russian intelligence agencies , citing a recent flareup of activity by \" Guccifer 2.0 , \" believed to be a Russia - sponsored puppet account , and the country 's well - known penchant for \" false flag \" operations , intended to deceive or confuse people .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceive\", \"contact.prevarication.broadcast\", \"An act of spreading misinformation to mislead others through deceptive communication channels.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThere is a medical document confirming the rape .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"rape\", \"life.injure.injurycausedbyviolentevents\", \"An injury resulting from violent acts, such as physical assault or rape, leading to physical or psychological harm to the victim.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIf the Clinton campaign , and not the Trump campaign , were so extensively interwoven with Putin ’s Russia , it ’s a safe bet that Trump would be demanding that Clinton release her tax returns to prove that she ’s not beholden to Putin — just as he demanded Obama release his birth certificate .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"demanded\", \"contact.commandorder.broadcast\", \"An entity issues a formal request or order for information or action, typically with authority or pressure.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n1 problem , then said it would be a good thing if we withdrew from Japan and South Korea , even if the two countries reacted by building their own nuclear arsenals .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"building\", \"manufacture.artifact.build\", \"The construction of a physical object or structure through manufacturing or assembly processes.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n\" It is impossible for the FBI not to recommend criminal charges against Hillary Clinton , \" Trump tweeted Saturday .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"recommend\", \"contact.requestadvise.correspondence\", \"An official request for advice or action through correspondence, indicating a formal suggestion made by one entity to another.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nFlames are pictured rising after the Russian army carried out airstrikes on two hospitals in Aleppo , Syria , last December\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"airstrikes\", \"conflict.attack.airstrikemissilestrike\", \"An aerial attack using missiles or bombs targeting a location during a conflict, causing destruction and harm to civilians or infrastructure.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTwenty - two voted in favour of the sanction , with one voting against .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"sanction\", \"transaction.transaction.embargosanction\", \"An embargo sanction is imposed on a country or entity, restricting trade or financial activities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nFeels he has the right to sexually assault women .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"assault\", \"life.injure.injurycausedbyviolentevents\", \"This event represents physical harm caused by violent actions, such as assault, involving direct physical injury to a person.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWe can keep going for two to three months , but then people will starve in large numbers .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"starve\", \"life.injure.illnessdegradationhungerthirst\", \"Hunger and thirst lead to physical deterioration and illness, resulting in widespread suffering and death among people over time.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nResponding to Orlando — the deadliest attack by a gunman in our history — the NRA blamed the massacre on \" the Obama administration 's political correctness \" and not the terrorist 's access to \" so - called ' assault weapons ' \" — while advertising that \" semi - automatics are the most popular firearms sold in America for sport - shooting , hunting and self - defense . \"\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"shooting\", \"conflict.attack.firearmattack\", \"An attack involving a firearm used to cause harm in a violent conflict, typically resulting in casualties during a mass shooting incident.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTrump also faces questions about what happened to money from his most recent book , “ Crippled America . ” He pledged to donate the profits to charity : “ In May , Trump reported earning more than $ 1 million in royalties from the book .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"pledged\", \"contact.commitmentpromiseexpressintent.meet\", \"An individual commits to donate profits from a book to charity, expressing intent to fulfill a financial pledge.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nFabius announced Friday that France will try to convene an international peace summit in the next few weeks to renew diplomatic efforts between Israeli and the Palestinians .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"convene\", \"contact.discussion.correspondence\", \"An event where parties engage in discussions or correspondence to establish or maintain communication for diplomatic or political purposes.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIsrael should stop building settlements , denying Palestinian development and designating land for exclusive Israeli use that Palestinians seek for a future state , the Middle East peace quartet – the US , Russia , the EU and the UN – recommended today in a an eagerly awaited report .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"recommended\", \"contact.requestadvise.broadcast\", \"An official recommendation made through a public channel by a group to influence policy or behavior regarding a specific issue.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn May 2015 , Ripley and his team were granted rare access to the Demilitarized Zone ( DMZ ) in North Korea .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"granted\", \"transaction.transaction.giftgrantprovideaid\", \"This event represents the granting of access, resources, or aid to an entity, typically involving permission or support.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIf crime rates remained the same ( which , of course , they did n’t ) , 83 percent of all Americans aged 12 at the time would , in their actuarial lifetimes , be victimized by an attempted or completed violent crime , and 40 percent would be injured as a result of a robbery or assault .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"assault\", \"life.injure.injurycausedbyviolentevents\", \"An individual suffers injury from a violent crime such as assault or robbery during their lifetime.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nSome onlookers have suggested that the alleged thieves could be linked to Russian intelligence agencies , citing a recent flareup of activity by \" Guccifer 2.0 , \" believed to be a Russia - sponsored puppet account , and the country 's well - known penchant for \" false flag \" operations , intended to deceive or confuse people .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceive\", \"contact.prevarication.meet\", \"An event where individuals meet to conceal information or mislead others through deceptive contact.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn May of last year , a Moroccan pilot died when his F-16 crashed in the north Yemen governorate of Saada .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"died\", \"life.die.nonviolentdeath\", \"An individual's death without violence, indicating natural or non-combat causes, distinct from violent or accidental fatalities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn June 1940 , German troops marched into Paris , forcing France to capitulate and establish the pro - Axis Vichy French government .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"capitulate\", \"conflict.yield.n/a\", \"An entity surrenders due to military pressure, resulting in a loss of sovereignty and the establishment of a hostile-aligned government.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nU.S. Secretary of State John Kerry told Turkish Foreign Minister Mevlut Cavusoglu by phone on Thursday that YPG fighters were retreating to the east side of the Euphrates , as Turkey has demanded , foreign ministry sources in Ankara said .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"retreating\", \"conflict.yield.retreat\", \"Fighters moving away from a conflict zone to a safer location as part of a strategic withdrawal or compliance with external demands.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n' During the drill we are going to inspect what was planned to be inspected and make suggestions on building - up the potential and areas of developing civil defense considering risks and threats , civil defense plans are going to be revised and suggestions will be made for summary rules of national standards . '\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"inspected\", \"inspection.sensoryobserve.physicalinvestigateinspect\", \"A physical examination of planned inspection items to assess conditions and evaluate civil defense readiness and risk exposure.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIsik said the retreat was not yet complete and Washington had given assurances that this would happen in the next week .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"retreat\", \"conflict.yield.n/a\", \"Indicates a reduction in conflict intensity or military engagement, suggesting a partial resolution or withdrawal from hostilities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe relentless airstrikes by the coalition , now in their 24th month , reflect the intensity of air campaign against ISIS , said Chris Woods , the director of Airwars , which is based in the United Kingdom .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"airstrikes\", \"conflict.attack.n/a\", \"Represents an attack conducted by military forces using airpower in a conflict context, indicating hostile actions against a target.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAlthough Democrats are complaining that the Project Veritas videos were edited ( as are virtually all TV videos ) , and therefore misleading , Creamer , as well as Foval , who admitted on tape to hiring “ mentally ill ” people to disrupt Trump events and foment violence , were ousted almost immediately , with Creamer stepping down and Foval being “ laid off . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"misleading\", \"contact.prevarication.correspondence\", \"This event captures false or deceptive communication in correspondence, indicating intentional distortion of facts in written or verbal exchanges.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nPhotos : World War II in pictures German prisoners captured at Friedrichsfeld march through a town in Germany after the crossing of the Rhine River by the U.S. 9th Army on March 26 , 1945 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"captured\", \"movement.transportperson.preventexit\", \"This event represents preventing a person from exiting during transport, indicating restricted movement of prisoners in military logistics.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ This boycott movement is enjoying growing success around the world as the only nonviolent means to put pressure on Israel , ” the public figures’ statement in the independent online publication Médiapart says .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"boycott\", \"transaction.transaction.embargosanction\", \"A global nonviolent boycott movement exerts economic pressure on Israel through consumer actions and international support.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ I do n’t see the point of going through the motions and creating a misleading impression that something else is going on , ” Cornyn said .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"misleading\", \"contact.prevarication.correspondence\", \"This event involves false or deceptive communication to misrepresent reality in official or public interactions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn June 1940 , German troops marched into Paris , forcing France to capitulate and establish the pro - Axis Vichy French government .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"capitulate\", \"conflict.yield.surrender\", \"A state where a party stops resistance and formally gives up, often due to military defeat or pressure in a conflict.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nShe sells to entry - level floor sweepers who make $ 8 an hour and ca n’t afford to get sick .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"sick\", \"life.injure.n/a\", \"An individual suffers a physical or health-related harm, often due to illness or injury, impacting their well-being or ability to function.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn her “ Fox News Sunday ” interview , Clinton was asked about Smith ’s allegations that she misled her and another Benghazi family after the 2012 attack .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"misled\", \"contact.prevarication.broadcast\", \"An instance where a public figure falsely claims to have misled others during a broadcast interview about sensitive political allegations.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMany of these countries include the world 's worst tyrants with abysmal human rights records and are known to fund terrorism , execute gays , and discriminate against women and religious minorities .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"execute\", \"justice.judicialconsequences.execute\", \"This event denotes the judicial or state-sanctioned killing of individuals, often linked to human rights violations and abuse of power.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nRUSSIAN President Vladimir Putin has taken a page out of Joseph Stalin ’s book — and sacked every commander in his Baltic fleet .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"sacked\", \"personnel.endposition.firinglayoff\", \"This event represents the removal of military commanders from their positions through dismissal or termination by a leader.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHere ’s something that does n’t work : inspecting a man ’s anus to determine if he ’s homosexual .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"inspecting\", \"inspection.sensoryobserve.n/a\", \"A sensory observation event involving physical examination to detect characteristics, such as sexual orientation, through direct inspection.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nSlaughter : No question .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"Slaughter\", \"conflict.attack.n/a\", \"Represents a violent attack with no specified category or context, indicating unclassified aggression or harm.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n\" Debbie Wasserman Schultz resigned over her failure to secure the DNC 's email servers and the rigged system she set up with the Clinton campaign , \" he said in a statement .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"resigned\", \"personnel.endposition.n/a\", \"An individual voluntarily leaves a position or office, indicating a termination of their role or responsibility in an organization.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nResponding to Orlando — the deadliest attack by a gunman in our history — the NRA blamed the massacre on \" the Obama administration 's political correctness \" and not the terrorist 's access to \" so - called ' assault weapons ' \" — while advertising that \" semi - automatics are the most popular firearms sold in America for sport - shooting , hunting and self - defense . \"\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"attack\", \"conflict.attack.n/a\", \"An attack event representing a violent act with no specific classification, indicating a conflict involving harm or aggression.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAnd of course , an army of government shills and trolls have thoroughly infiltrated if not saturated both mainstream as well as alternative internet and social media with the expressed purpose of muddying the waters to intentionally misinform , deceive and confuse the public , making them highly susceptible to nefarious manipulation by fear porn propaganda .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceive\", \"contact.prevarication.broadcast\", \"An act of spreading false information to mislead public perception and manipulate opinions through deceptive communication channels.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n[ Putin implicated in fatal poisoning of former KGB officer at London hotel ]\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"fatal\", \"life.die.deathcausedbyviolentevents\", \"An individual dies due to violent actions, indicating a lethal outcome from intentional or hostile physical intervention.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAfter the November Paris attack , I wrote , “ ISIS certainly intends to strike the U.S. homeland , so unless it is annihilated , it ’s just a matter of time before it will be able to carry out a domestic attack of the kind it recently executed in Paris , Beirut and aboard a Russian jetliner . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"annihilated\", \"conflict.attack.n/a\", \"An unspecified attack event indicating intent to strike a target, often implying threat or potential violence against a nation or group.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nPence confirmed that Trump is scheduled to make a speech Monday in battleground state Ohio on the issue and vowed that Trump would “ lay out his vision and his strategy for defeating radical Islamic terrorism . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"vowed\", \"contact.commitmentpromiseexpressintent.broadcast\", \"An official declaration where a person commits to expressing a political vision or strategy through a public broadcast.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn The Code of the Streets , Elijah Anderson analyzed the reality that many black people in violent neighborhoods live under and the grim rules ( \" the code \" ) they ’ve adopted that essentially allow violence to solve disputes the law ca n’t be trusted to take care of .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"violent\", \"life.injure.n/a\", \"This event signifies physical harm or injury occurring in a violent context within a community setting.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWada has recommended the IOC \" decline entries , for Rio 2016 , of all athletes \" submitted by the Russian Olympic Committee and the Russian Paralympic Committee .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"recommended\", \"contact.requestadvise.correspondence\", \"An official recommendation made through correspondence to influence decisions regarding athlete participation in an Olympic event.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nShe will no doubt continue to defend the system heart and soul , as will her husband , Bill Clinton , who was known in his time as the \" globalization president . ” ( “ There is no longer a clear division between what is foreign and what is domestic , \" he said at his first inaugural in 1993 , and reiterated the point in his final foreign policy address in 2000 ; both sentiments now appear to be outdated in this time of Trump . )\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"reiterated\", \"contact.publicstatementinperson.n/a\", \"An in-person public statement reaffirming a previous position or policy view during a formal address or speech.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nFox host Bill O'Reilly asked Trump how to solve the city 's crime problem , with homicides up nearly 50 percent .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"homicides\", \"life.die.deathcausedbyviolentevents\", \"Death resulting from violent acts such as shootings or assaults, indicating a significant increase in fatal crimes in a city.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBlack Lives Matter demonstrators march in Cleveland on December 29 , 2015 , after a grand jury declined to indict Cleveland Police officer Timothy Loehmann for the fatal shooting of Tamir Rice on November 22 , 2014 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"fatal\", \"life.die.deathcausedbyviolentevents\", \"This event type represents a death resulting from violent actions, indicating a fatal outcome due to physical aggression or force.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n\" Authorities reported no fatalities , but crews had shifted into search mode Thursday morning , seeking to account for five to seven residents of the two adjacent buildings at the Flower Branch Apartments on Piney Branch Road .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"fatalities\", \"life.die.deathcausedbyviolentevents\", \"This event indicates deaths resulting from violent incidents, signaling loss of life due to conflict or aggression in a specific location and time.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBillionaire tech investor Peter Thiel donated $ 2 million to a group supporting Carly Fiorina .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"donated\", \"transaction.transaction.giftgrantprovideaid\", \"An individual gives money to a group for support, indicating financial aid or charitable contribution.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWe can keep going for two to three months , but then people will starve in large numbers Brita Haji Hasan\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"starve\", \"life.injure.illnessdegradationhungerthirst\", \"Individuals suffer physical decline due to prolonged lack of food and water, leading to severe health deterioration and death.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTrump , in condemning one of the worst instances of American overreach in U.S. history , the Iraq invasion , declared in his speech : “ The world must know we do not go abroad in search of enemies . ” The line was an allusion to the famous injunction of John Quincy Adams in 1821 that America \" does not go in search of monsters to destroy .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"injunction\", \"transaction.transferownership.embargosanction\", \"This event represents a transfer of ownership or sanction imposed through an embargo, restricting trade or actions against a nation or entity.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe candidate ’s real estate empire , unable to borrow from most U.S. banks , gets capital from Russian sources .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"borrow\", \"transaction.transferownership.n/a\", \"This event represents a transfer of ownership without financial borrowing, indicating a shift in control or asset management.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMore than 40 percent of Russian families struggle to find the money to buy food or clothes , a study by Moscow ’s Higher School of Economics ( HSE ) revealed Wednesday .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"buy\", \"transaction.transferownership.n/a\", \"Represents a transfer of ownership in a transaction, where one party gains control over an item or asset through purchase or exchange.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe mix of donations also exposed a seemingly common practice of wealthy individuals donating at the request of clients and friends .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"donating\", \"transaction.transfermoney.giftgrantprovideaid\", \"Represents the act of transferring money as a gift, grant, or form of aid without commercial intent.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ The review will be very much exact and recommended by experts , ” Trump added .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"recommended\", \"contact.requestadvise.n/a\", \"An expert provides advice or endorsement for a review, indicating professional validation and recommendation.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTrump , in condemning one of the worst instances of American overreach in U.S. history , the Iraq invasion , declared in his speech : “ The world must know we do not go abroad in search of enemies . ” The line was an allusion to the famous injunction of John Quincy Adams in 1821 that America \" does not go in search of monsters to destroy .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"injunction\", \"transaction.transfermoney.embargosanction\", \"An economic sanction imposed by a state to restrict financial transactions with another entity as a form of political or diplomatic pressure.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n— His employees violated a U.S. trade embargo by pursuing commercial activities in Cuba when it was illegal to do so\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"embargo\", \"transaction.transferownership.embargosanction\", \"Transfer of ownership occurred in violation of U.S. trade embargos, specifically prohibited commercial activities in Cuba.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe U.N. Security Council has long threatened to impose an arms embargo on South Sudan but veto - power Russia has been reluctant to support such an action .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"embargo\", \"transaction.transfermoney.embargosanction\", \"An embargo sanctions the transfer of goods or weapons to a country, restricting trade or military support.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe embargo , restricting imports of most food from the European Union , the United States , Canada , Australia and Norway , was instituted in 2014 in response to economic sanctions placed on Russia for its military and political involvement in Ukraine .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"embargo\", \"transaction.transferownership.embargosanction\", \"An embargo restricts trade or imports, typically imposed due to political or economic sanctions against a country or entity.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nOn Saturday a great man , Elie Wiesel , passed away . The world is a better place because of him and his belief that good can triumph over evil !\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"passed away\", \"life.die.nonviolentdeath\", \"A non-violent death of a notable individual, emphasizing peaceful passing without physical harm or conflict.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\ntwitter email Republican presidential nominee Donald Trump urges Russia to find his Democratic rival Hillary Clinton 's emails during a news conference on Wednesday , July 27 , 2016 , at Trump National Doral .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"urges\", \"contact.commandorder.broadcast\", \"An official directive sent publicly to influence another party to take a specific action, such as seeking information.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n\" We have everything at the table , we are prepared to look at sanctions , we 're prepared to look at an arms embargo , \" U.S. Special Envoy to Sudan and South Sudan , Donald Booth , told lawmakers .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"embargo\", \"transaction.transfermoney.embargosanction\", \"An event indicating financial or trade restrictions imposed through sanctions or embargos on a country or entity.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nA neo - Nazi group , including former Latvian Nazi Waffen SS veterans , has demolished a monument to fallen Soviet troops in a Latvian town with the support of a nationalist party which forms part of the ruling coalition in Latvia ’s parliament .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"demolished\", \"artifactexistence.damagedestroy.n/a\", \"An artifact's existence is damaged or destroyed, indicating physical removal or destruction of a structure or object.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n\" The milk market remains extremely difficult ; Russia continues its trade sanctions against the EU .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"sanctions\", \"transaction.transaction.embargosanction\", \"An embargo sanction restricts trade between parties, often due to political or economic disagreements, impacting market access and economic flows.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe State Department has since resumed its investigation into whether Mrs Clinton or her aides violated government policy in their handling of classified information .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"investigation\", \"justice.investigate.investigatecrime\", \"An official inquiry into potential policy violations involving classified information handling by individuals in government roles.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn 2015 Russian president Vladimir Putin extended the embargo through Aug. 5 , 2016 and added four more countries to the list : Albania , Montenegro , Iceland and Liechtenstein .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"embargo\", \"transaction.transfermoney.embargosanction\", \"An embargo sanction extends restrictions on trade or financial transactions to specific countries or entities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nComey on July 5 said he would n’t recommend charges against Clinton for her use of a private e - mail server while secretary of state .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"recommend\", \"contact.requestadvise.broadcast\", \"An official requests advice or recommendation regarding a matter through a public or formal channel.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nRoosevelt died on April 12 , 1945 , in Warm Springs , Georgia , just weeks before Germany 's surrender .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"died\", \"life.die.nonviolentdeath\", \"An individual died without violence or conflict, typically due to natural causes, in a peaceful or non-combat setting.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nPerhaps the most famous example of this skill came when Mr. Trump was asked under oath a seemingly straightforward question : Had he ever lied about his net worth ?\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"oath\", \"contact.commitmentpromiseexpressintent.broadcast\", \"An official declaration made under legal obligation, expressing intent or commitment through formal communication.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe second explosion was a sucicde bomber who detonated his belt as people rushed to help those injured , AFP reported .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"detonated\", \"conflict.attack.n/a\", \"An unclassified attack event involving detonation, indicating a violent act with no specific conflict classification.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAs the citizens are promised ‘ reform ’ the only thing that actually changes are the puppets in marble buildings .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"promised\", \"contact.commitmentpromiseexpressintent.broadcast\", \"An entity publicly commits to a change or reform through a promise, signaling intent to act or modify policies.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nKORNACKI : In terms of the political battle that we 're seeing play out right now , the White House is saying the Republicans have no business obstructing a nominee the president puts forward .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"obstructing\", \"movement.transportartifact.prevententry\", \"Preventing an artifact from entering a designated area or space through active resistance or interference.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAn estimated 60,000 to 70,000 were killed in the Nagasaki blast .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"blast\", \"conflict.attack.bombing\", \"A bombing attack during a conflict involving the use of explosives to cause harm or destruction in a military or civilian context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAleppo is therefore key to the strategy pursued by Russia , which will not stop bombing it for the sake of the Vienna process created by Russia or for fear of European or American reactions .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"bombing\", \"conflict.attack.n/a\", \"An attack involving bombing, indicating hostile military action against a target in a conflict context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nOn the U.S. Federal Reserve , Trump said that while he eventually wants a Republican to head it , he is \" not an enemy \" of current chair Janet Yellen , who was appointed by Obama .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"appointed\", \"personnel.startposition.hiring\", \"This event signifies the formal appointment of a person to a new position within an organization or government body.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBut because of Edward Snowden , we now know the head of US intelligence was lying to the Senate committee responsible for intelligence oversight .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"lying\", \"contact.prevarication.meet\", \"An intelligence official concealed truth from a Senate committee during a meeting, violating transparency and oversight protocols.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nDue to the Clinton - instigated settlement , UBS only had to disclose information on 4,450 accounts to the IRS held by Americans , which ameliorated at least some of the tension UBS faced — namely violating Swiss secrecy laws or staring down U.S. criminal charges .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"charges\", \"justice.initiatejudicialprocess.chargeindict\", \"The event represents the initiation of legal charges against an entity for alleged violations of laws or regulations, leading to judicial proceedings.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe over - riding mission of the USOC is winning the most medals possible ; when Blackmun discussed how his organization decides which sports and athletes deserve money , he used the phrase “ podium potential . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"discussed\", \"contact.discussion.meet\", \"An event where two parties engage in a conversation to exchange ideas or make decisions about organizational priorities or policies.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ We got funding for the film but Paul ’s company , Manhattan Pictures , ran out of money and shuttered .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"funding\", \"transaction.transfermoney.giftgrantprovideaid\", \"This event represents financial support provided to an entity for a specific purpose, such as funding a film project.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBill O'Keefe , Catholic Relief Service 's vice president for advocacy , said in an interview that while his group and others welcomed the administration 's new resettlement pledge , neither that nor the upcoming U.N. summit was sufficient to address the needs of \" the 65 million globally displaced people around the world right now . \"\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"pledge\", \"contact.commitmentpromiseexpressintent.meet\", \"An entity expresses an intention to commit or promise support or action through a formal commitment during a meeting or interaction.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe CNAS paper urges Congress to pass the TPP as soon as possible , presumably before the election , to strengthen the U.S. “ stewardship ” of Asia .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"urges\", \"contact.requestadvise.meet\", \"An official requests advice or a meeting to influence policy, typically to advocate for a specific action or decision.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMr Roslovtsev says he is protesting over the exact laws which are cited in his arrests , which he describes as \" absurd \" .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"protesting\", \"conflict.demonstrate.marchprotestpoliticalgathering\", \"A public demonstration or protest against political policies or laws, reflecting collective opposition to specific government actions or regulations.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe best way to start on this path is to repudiate the candidate before the election and avoid going down with the USS Trump .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"election\", \"inspection.sensoryobserve.monitorelection\", \"Monitoring an election involves observing electoral processes to ensure fairness, detect fraud, or assess public sentiment during voting periods.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWill your parents have to borrow money to bury you or your wife or girlfriend ?\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"borrow\", \"transaction.transferownership.borrowlend\", \"An event where one party borrows or lends money, involving a transfer of financial ownership between individuals.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nNothing was taken , but police have said attempted robbery is their leading theory for a motive , noting a spike in robberies in the neighbourhood in the preceding weeks .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"robberies\", \"conflict.attack.stealrobhijack\", \"An event where individuals attempt to steal property through force or threat, commonly involving armed confrontation or illegal intrusion.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWith his family surrounding him on stage , Cruz said he would end his campaign but vowed that the conservative movement he leads would go on .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"vowed\", \"contact.commitmentpromiseexpressintent.broadcast\", \"An individual publicly commits to a future action or belief, expressing intent through a formal declaration in a public setting.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAnd for \" The Voting Wars , \" I was able to find regularly in local elections - and certainly in some parts of the country there 's been an unfortunate history of this , in parts of Kentucky , South Texas , parts of Florida .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"Voting\", \"government.vote.n/a\", \"This event denotes voting activity in local elections, reflecting participation or disputes in electoral processes.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nCruz accused his leader of deceiving him about whether he made a secret deal with Democrats to allow a vote reauthorizing the Export - Import Bank .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceiving\", \"contact.prevarication.meet\", \"An encounter where one party withholds or misrepresents information to another, involving deception in a meeting context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n-- The Khan family is becoming a big issue down the ballot -- > “ Khizr Khan calls on John McCain to stand up to Trump , ” by Cronkite News ’s Emily Zentner ( they are affiliated with Arizona PBS ) : “ The last book Khizr Khan sent his son Humayun , an Army captain who was killed in a 2004 car bombing in Iraq , was Sen. John McCain ’s ‘ Why Courage Matters . ’ Khan said he has long respected the Arizona Republican , and that one of his last conversations with his son had been about the book and McCain ’s sacrifice … McCain had been ‘ my and my family ’s hero , ’ whom Khan said they admired as someone who ‘ gave so much in care of others . ’ ... It ’s that admiration that led Khan to call on McCain and other GOP leaders to withdraw their support of Trump : ‘ I implore Sen. McCain … this will be a historic moment in the Republican Party , ’ Khan said during an interview with Cronkite News .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"killed\", \"life.die.deathcausedbyviolentevents\", \"An individual dies as a result of violent events, indicating a fatal outcome due to conflict, attack, or combat-related causes.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nProsecutors blamed witness tampering and obstruction by Kenya ’s government for the collapse of that case .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"obstruction\", \"movement.transportperson.prevententry\", \"Preventing a person from entering a location through illegal or obstructive actions, often involving interference with movement or access.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHe has claimed that undocumented immigrants are “ taking our jobs ” and “ taking our money , ” pledged to deport them en masse , and vowed to build a wall on the Mexican border .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"vowed\", \"contact.commitmentpromiseexpressintent.meet\", \"An individual expresses a firm intention to take specific actions regarding undocumented immigrants through public commitment or pledge.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe DNC called in computer security company CrowdStrike during the first week of May to investigate the suspected hack .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"investigate\", \"justice.investigate.investigatecrime\", \"An official inquiry into a crime or suspicious activity, initiated by authorities to determine the cause and perpetrators.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nFirst , the decision to minimize her chronic dishonesty baffles me because if you elect a president you do n’t trust , you forfeit any right to complain when you are deceived .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceived\", \"contact.prevarication.n/a\", \"This event captures a situation where someone is misled due to dishonesty, indicating a breach of trust in interpersonal or political interactions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n( It ’s not just conservatives , either ; this is the kind of logic that led Hillary Clinton to speak out against violent teenage \" superpredators \" in the 1990s , when her husband , as president , enacted \" tough - on - crime \" laws . )\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"violent\", \"life.injure.n/a\", \"Represents an injury to a living being, indicating harm or damage to physical or mental health without specifying the cause or mechanism.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nFalsely accused Clinton of laughing at a 12-year - old rape victim : “ Her client she represented got him off and she 's seen laughing at the girl who was raped . ” ( Clinton laughed at various points of a 1980s interview about the case – but never at the victim .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"rape\", \"life.injure.injurycausedbyviolentevents\", \"An injury resulting from violent events, such as physical or psychological harm caused by aggression or abuse.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAlexander Sokolov and Dorothee Ehrich said they found 15 patches of ground that appeared to bubble or tremble and when they punctured them , methane and carbon dioxide was released , the Siberian Times reports .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"punctured\", \"artifactexistence.damagedestroy.n/a\", \"This event indicates the act of damaging or breaking an object, specifically by piercing it to release internal contents.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThis comes just as the French prime minister has announced that his government plans to intensify its restrictions on free speech targeting the boycott , divestment and sanctions ( BDS ) movement .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"boycott\", \"transaction.transaction.embargosanction\", \"An official government action restricting activities against a movement, specifically targeting BDS through speech limitations.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIt also contains a link to the genocide - denial group FactCheckArmenia.com , which propagates claims that Armenians were killed during the First World War because they were “ collectively guilty ” of treason against the Ottoman Empire .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"guilty\", \"justice.judicialconsequences.convict\", \"This event indicates a judicial finding of guilt against an individual or group for a criminal offense involving wrongdoing or misconduct.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe current DNC chair is now embroiled in controversy and one poll shows 31 % of Democrats wanted the FBI to recommend indictments after Clinton ’s email probe .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"recommend\", \"contact.requestadvise.correspondence\", \"An entity requests advice or suggestion through correspondence, often involving official recommendations or guidance on a specific matter.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nLord Coe and council convened to decide whether to suspend Russian athletes from competition\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"convened\", \"contact.discussion.meet\", \"A meeting between individuals to discuss a decision regarding athlete participation in competition.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn exchange for Iran ’s adherence to these terms , the UN Security Council members have agreed to reduce sanctions on the Islamic Republic .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"sanctions\", \"transaction.transaction.embargosanction\", \"Reduction of economic sanctions imposed on a country in exchange for compliance with international conditions or agreements.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe deal , which required that Assad halt airstrikes against civilian areas in places like Aleppo , quickly collapsed and the U.S. and Russia have cut off formal dialogue about Syria .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"airstrikes\", \"conflict.attack.airstrikemissilestrike\", \"Airstrikes involve military air attacks on targets, often targeting civilian areas during conflicts, causing harm and disrupting peace.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe New York Times reported that Mr Powell had once advised Mrs Clinton at a dinner party to use private email , although not while handling classified information .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"advised\", \"contact.requestadvise.n/a\", \"An event where one person advises another on a specific action or behavior during a personal interaction.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBut in practice , the government has taken back a number of areas with starve - or - surrender tactics , bombarding and starving people until they agree to leave .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"bombarding\", \"conflict.attack.n/a\", \"An attack conducted during conflict, involving force or violence against a target group or region in a hostile situation.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIt calls on the WHO to convene an independent group to advise it and the International Olympic Committee , and for authorities to reconsider the decision to hold the Games in Rio .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"convene\", \"contact.discussion.meet\", \"An event where parties come together to discuss issues, involving dialogue or consultation between organizations or individuals.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nA world existed where the arc of Weiner 's career looked like this : sext while a member of Congress , resign , apologize , run for mayor , weather revelations that he did n't stop sexting when he got caught the first time , lose the primary for mayor , star in an unflinching documentary about that painfully embarrassing moment AND possibly have an opening for a third chance at public office .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"resign\", \"personnel.endposition.n/a\", \"This event signifies the termination of a person's official position in an organization or office, such as leaving public office.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n\" Mr. Murphy has also challenged Mr. Rubio ’s obstinate support for the failed embargo on Cuba , which puts him on the right side of history and , increasingly , public opinion in Florida , \" it stated .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"embargo\", \"transaction.transfermoney.embargosanction\", \"An event where financial or political sanctions are imposed through an embargo on a country, restricting trade or aid.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAkin ’s comments on “ legitimate rape ” and Mourdock ’s claim that pregnancy resulting from rape “ is something that God intended ” not only sank their own campaigns but also became political liabilities for candidates in other states by alienating women .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"rape\", \"life.injure.n/a\", \"This event type signifies a harm or injury to a person's life, often involving physical or psychological damage from violence or abuse.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMax Boot , the Commentary columnist and senior fellow at the Council on Foreign Relations who ’d been advising Marco Rubio before he dropped out , told the New York Times that he ’d vote for Josef Stalin over Trump .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"advising\", \"contact.requestadvise.n/a\", \"An individual seeks advice from another in a professional or strategic capacity, indicating a relationship of guidance and influence.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nDemocratic Nominee for Vice President Tim Kaine walks on stage during the 2016 Democratic National Convention at Wells Fargo Center .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"walks\", \"movement.transportperson.selfmotion\", \"Represents a person moving on their own without external transport, indicating physical movement from one location to another.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nPhotos : World War II in pictures Balloon operators from Britain 's Women 's Auxiliary Air Force , or WAAF , report for inspection in a hangar used to store balloons , at a facility in the UK .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"inspection\", \"inspection.sensoryobserve.inspectpeopleorganization\", \"An official check of personnel or groups using sensory observation in a structured organizational setting.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAnd the thing about a word like fascism — not unlike genocide or rape or cancer — is that , once it is spoken , it can not be allowed to proceed .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"genocide\", \"life.die.deathcausedbyviolentevents\", \"Death resulting from violent, large-scale human actions such as genocide, involving intentional mass harm and systemic destruction.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n\" The ' inghimasi , ' their ( modus operandi ) on the ground in Syria and Iraq , is to shoot up checkpoints and then they actually -- some of these guys actually run up to the enemy and hug them before detonating the bomb to take them out with themselves .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"detonating\", \"conflict.attack.bombing\", \"An attack involving the use of a bomb to destroy targets, typically in a conflict zone, resulting in harm or death to enemies or civilians.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nEvery once in a while you hear a story , so you hear a story about a husband passes away , gets an absentee ballot in the mail , widow votes for husband .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"passes away\", \"life.die.nonviolentdeath\", \"A person dies without violence, typically due to natural causes, marking the end of life in a peaceful or non-aggressive manner.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAccording to her lawyers , she was captured by the rebels on the morning of 17 June whilst trying to reach several Aidar fighters who had been wounded in an ambush .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"captured\", \"justice.arrestjaildetain.arrestjaildetain\", \"An individual is taken into custody by authorities or hostile forces, often involving unlawful detention or capture.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAll eyes were turning to Rio on Sunday as the IOC convened to consider throwing Russia out of the Games .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"convened\", \"contact.discussion.meet\", \"A meeting where parties discuss issues, often involving decisions or evaluations regarding policies, actions, or participants in a structured setting.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe experiences the former Trump models related to Mother Jones echo allegations in an ongoing class - action lawsuit against six major modeling agencies by nine former models who have claimed their agencies charged them exorbitant fees for rent and other expenses .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"rent\", \"transaction.transferownership.borrowlend\", \"Represents a financial exchange where one party lends or borrows assets, including rent-based transactions between models and agencies.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe shield , which is modeled on the United States ' so - called Aegis ships , was assembled in New Jersey and then transferred to the Deveselu base in containers .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"assembled\", \"manufacture.artifact.build\", \"The construction of a physical object through manufacturing processes involving assembly and integration of components into a functional artifact.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nLast month , Trump caused a stir by refusing to unconditionally pledge to come to the aide of NATO allies in Europe were they to be invaded by Russia .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"pledge\", \"contact.commitmentpromiseexpressintent.meet\", \"An event where an entity expresses an intent to commit or promise support to another entity under specific conditions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAt the last UEFA European Championship , played in Poland in 2012 , Russia was penalized for the poor behavior of traveling supporters -- including violent disorder and the racial abuse of a black player from Czech Republic\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"violent\", \"life.injure.injurycausedbyviolentevents\", \"An injury resulting from violent behavior by supporters during a sports event, involving physical or emotional harm to individuals.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe Kremlin launched its air strikes saying it wanted to help Assad , its main Middle East ally , defeat Islamic State and other militant groups .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"strikes\", \"conflict.attack.bombing\", \"An aerial bombing campaign conducted by a state actor to target militant groups in a conflict zone, often as part of a broader military strategy.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThey must recreate in the public mind that Hillary Clinton , who 56 percent of the nation thought should have been indicted for lying in the server scandal and who two - thirds of the nation said was dishonest or untrustworthy .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"indicted\", \"justice.initiatejudicialprocess.chargeindict\", \"This event represents the initiation of a judicial process to formally charge an individual with a criminal offense based on evidence of wrongdoing.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThis week Brazil moved one step closer to removing Dilma Rousseff from office after the Senate voted to indict the suspended president over charges of violating fiscal laws .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"charges\", \"justice.initiatejudicialprocess.chargeindict\", \"The event signifies the formal accusation of a public official for violating fiscal laws through a judicial indictment process.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nOne victim said under oath that Epstein molested her at least 50 times , beginning when she was 13 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"oath\", \"contact.commitmentpromiseexpressintent.meet\", \"An event where one party expresses intent to meet another through a formal declaration under oath.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTrump , in condemning one of the worst instances of American overreach in U.S. history , the Iraq invasion , declared in his speech : “ The world must know we do not go abroad in search of enemies . ” The line was an allusion to the famous injunction of John Quincy Adams in 1821 that America \" does not go in search of monsters to destroy .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"injunction\", \"transaction.transaction.embargosanction\", \"An official restriction imposed by a government on trade or activities with a country or entity, often due to political or economic reasons.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nResponding to a lawmaker 's question if the military could return to Cuba and Vietnam , Pankov said the military is \" reviewing \" a decision to withdraw from them , but did n't offer any specifics .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"withdraw\", \"conflict.yield.retreat\", \"A military retreat from a conflict zone indicates a strategic withdrawal, often due to changing circumstances or reduced operational capacity.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nPublic figures or advocates demonize a person or group as a loathsome , inhuman monster , often using violent imagery and righteous rhetoric — like comparing abortion to slavery , or saying it ’s worse than the Holocaust .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"Holocaust\", \"life.die.deathcausedbyviolentevents\", \"Death resulting from violent events, indicating harm caused by extreme or violent actions within a societal or historical context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n\" I remember getting whipped so hard I did n't know if I was going to survive .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"whipped\", \"life.injure.n/a\", \"An event where a person suffers physical harm due to being whipped, indicating severe injury or pain in a personal experience.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTrump 's candidacy is built on his business \" credibility . \"\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"built\", \"manufacture.artifact.n/a\", \"This event type represents the creation or establishment of something based on perceived credibility or trustworthiness in a given context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMost modern internationalists , both Democratic and Republican , have long since relegated John Quincy ’s injunction to history , that of a 19th century America that was still a developing country and wanted only to be left alone .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"injunction\", \"transaction.transaction.embargosanction\", \"An official restriction imposed by a government on trade or diplomatic relations with another country, often due to political or economic reasons.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nRobert S. Ford , a former ambassador to Syria , said , “ Many people working on Syria for the State Department have long urged a tougher policy with the Assad government as a means of facilitating arrival at a negotiated political deal to set up a new Syrian government . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"negotiated\", \"contact.negotiate.correspondence\", \"An official communication between parties to reach a mutual agreement on policy or governance through diplomatic discussion.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTrump ’s remarks are artfully deceiving .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceiving\", \"contact.prevarication.correspondence\", \"This event involves deceptive communication where one party intentionally misleads another through false or misleading statements.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nLaotian President Bounnhang Vorachit and President Barack Obama inspect an honor guard at the Presidential Palace in Vientiane , Laos on Tuesday .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"inspect\", \"inspection.sensoryobserve.inspectpeopleorganization\", \"An official checks a group of people or organization as part of a ceremonial or formal observation activity.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIt ’s absolutely perplexing to me to try to understand people involved in The Bilderberg Group , or see people , even friends and colleagues in places such as Los Angeles , spending hours and hours and hundreds of dollars bleaching their hair every week and purchasing absurdly priced designer clothing whilst in travels on the same tours through Europe and witness to young Syrian refugees or young disenfranchised children and kids in Detroit or Mexico City or even London on the streets , starved .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"purchasing\", \"transaction.transferownership.n/a\", \"Represents a transfer of ownership through purchase, indicating a financial transaction where value is exchanged for goods or services.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWith the Teapublican right wing victories of 2010 and 2014 , the corporate Democrats under Obama helped create their own excuse for failing to advance a progressive agenda : newly empowered Republican “ obstruction ” in Congress .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"obstruction\", \"movement.transportperson.prevententry\", \"Preventing a person from entering a location through transportation-related actions, typically involving barriers or restrictions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nShe pushed the United States to invade Libya and destabilize it , turning it into a hellhole where a U.S. ambassador and three other Americans were killed in an organized terror attack she suggested should be blamed on a YouTube video .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"invade\", \"conflict.attack.n/a\", \"Represents an attack event with no specific classification, indicating a violent conflict initiated or supported by a political or strategic actor.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMr. Blair has said that he had suggested that she get into the commodities market , and that he used his knowledge of trading to guide her along the way .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"suggested\", \"contact.requestadvise.correspondence\", \"An individual proposes advice or guidance to another through communication, indicating a request for expert input or direction.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nClinton suspended her campaign Tuesday to attend the funeral in Rhode Island of her friend Mark Weiner , a Democratic party donor and fundraiser .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"funeral\", \"contact.funeralvigil.meet\", \"An event where individuals gather for a funeral service to honor and pay respects to a deceased person.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nInside the siege barricades , protesters gathered in the neighbourhood of Shaar on Thursday night with banners reading ‘ Revolutionaries do n’t give up ’ , ‘ No to displacement ’ , and ‘ No to Putin ’ to demand a permanent end to the shelling , rather than just a ceasefire .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"demand\", \"contact.commandorder.n/a\", \"This event represents a call for action, where protesters issue a formal request for policy change or cessation of hostilities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHis family , Greek , fled Smyrna when Turks invaded this borderland in around 1920 , and he wound up studying engineering at Battersea Polytechnic .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"invaded\", \"conflict.attack.invade\", \"An invasion event where one group enters and attacks another in a borderland region, often leading to displacement and conflict.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIt ’s absolutely perplexing to me to try to understand people involved in The Bilderberg Group , or see people , even friends and colleagues in places such as Los Angeles , spending hours and hours and hundreds of dollars bleaching their hair every week and purchasing absurdly priced designer clothing whilst in travels on the same tours through Europe and witness to young Syrian refugees or young disenfranchised children and kids in Detroit or Mexico City or even London on the streets , starved .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"starved\", \"life.injure.n/a\", \"This event denotes a state of physical or existential harm due to deprivation, often linked to poverty or lack of basic necessities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn comparison , Chicago recorded 473 homicides for all of 2015 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"homicides\", \"life.die.deathcausedbyviolentevents\", \"This event represents deaths resulting from violent acts, specifically homicides, indicating fatal violent conflicts in a given period.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe embargo , restricting imports of most food from the European Union , the United States , Canada , Australia and Norway , was instituted in 2014 in response to economic sanctions placed on Russia for its military and political involvement in Ukraine .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"embargo\", \"transaction.transaction.embargosanction\", \"An embargo restricts trade or imports, typically imposed due to political or economic reasons involving sanctioned countries or regions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ There ’s very little incentive to do so since no one ever gets prosecuted for violating the rules . ” She said she ’s been shopping a package of tougher lobby disclosure rules drafted by an American Bar Association panel — the last full reform took place in the mid-1990s , after more than 50 years without any significant revision — but “ there ’s very little appetite for reform . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"violating\", \"government.agreements.violateagreement\", \"Violation of a government agreement occurs when parties fail to comply with established rules or commitments set by authorities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nOne former female inmate , who declined to be named because of the social stigma attached to drug offenders in Thailand , said she was jailed for three years after being caught with 20 methamphetamine pills .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"jailed\", \"justice.arrestjaildetain.arrestjaildetain\", \"An individual is arrested and detained in a legal system for violating drug-related laws, resulting in incarceration.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ We were sacked because the new government needs scientists who will applaud increased logging , to convince public opinion that this insane idea is okay , ” said Przemysław Chylarecki , one of the dismissed scientists .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"sacked\", \"personnel.endposition.firinglayoff\", \"An employee is removed from their job position due to organizational decisions or policy changes, often resulting in loss of employment or role.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n\" Russia 's clear stand on the need to combat terrorism mirrors our own , \" Modi said .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"combat\", \"conflict.attack.n/a\", \"This event represents an action to confront or eliminate threats through force or measures, indicating a stance against hostile acts.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe hospital attacks come as airstrikes by the United States and France inside Syria have killed as many as 140 civilians in a northwestern city and on its outskirts over the past couple of days .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"airstrikes\", \"conflict.attack.n/a\", \"Represents an attack incident during a conflict, involving violence against civilians or infrastructure in a geopolitical setting.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIf crime rates remained the same ( which , of course , they did n’t ) , 83 percent of all Americans aged 12 at the time would , in their actuarial lifetimes , be victimized by an attempted or completed violent crime , and 40 percent would be injured as a result of a robbery or assault .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"injured\", \"life.injure.injurycausedbyviolentevents\", \"An individual suffers physical harm due to violent criminal acts like robbery or assault during their lifetime.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAbandoning the Iranian nuclear agreement and reintroducing sanctions ?\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"sanctions\", \"transaction.transaction.embargosanction\", \"Imposing trade restrictions on a country to pressure it into changing policies or behavior, often related to nuclear programs or international agreements.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ The regime is using starvation as a weapon and the west does not seem bothered by it .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"starvation\", \"life.injure.n/a\", \"This event represents harm to living beings through deprivation, specifically starvation, leading to physical or health-related injury.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAmnesty 's crisis response director , Tirana Hassan , said : \" Turkey 's highly selective practice is appalling - only severely injured people are allowed entry to seek medical treatment while everyone else fleeing the violence is left unprotected . \"\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"injured\", \"life.injure.n/a\", \"This event represents a situation where individuals suffer physical injury, often due to violence or conflict, with no specific action or cause defined.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHe is believed to have died a few hours earlier , between 8 and 9.30pm .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"died\", \"life.die.deathcausedbyviolentevents\", \"This event indicates a death resulting from violent or aggressive actions, involving harm caused by physical force or conflict.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWhether vindicating war and torture and civil liberties violations , talking past the War on Drugs and the carceral state , or exhibiting coolness toward the type of popular protest seen during of Occupy Wall Street , with its direct attacks on a sort of American Sonderweg , establishment Democrats are adept at using a more “ realistic ” brand of Americentrism to consolidate power and anchor the party in the status quo .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"torture\", \"life.injure.injurycausedbyviolentevents\", \"This event represents harm caused by violent actions, specifically involving physical or psychological injury from torture or extreme violence.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nOfficer Jason Van Dyke was charged with murder in November in the fatal shooting of 17-year - old Laquan McDonald , and just last week the Police Department announced plans to seek the firing of seven officers for allegedly lying about the shooting .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"fatal\", \"life.die.deathcausedbyviolentevents\", \"This event represents a death resulting from violent actions, specifically a fatal shooting incident involving a young individual.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe Trump acolytes claim it does n’t matter ; he can hire experts to advise him .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"advise\", \"contact.requestadvise.meet\", \"An event where one party requests expert advice for decision-making or guidance in a specific context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nDid the president not lay the groundwork for the Republicans obstructing him on this ? \"\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"obstructing\", \"movement.transportartifact.prevententry\", \"Preventing an artifact from entering a location through transportation, blocking access or movement.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWe do n't need Bible reeducation , we need a helping hand out of the mess so that we can build a solid support system to help the children and ex - members . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"build\", \"manufacture.artifact.n/a\", \"This event type refers to the creation of a physical object or structure to support a social or community need.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTrump later barred the Washington Post from covering his rallies after the newspaper ran an article about those comments with the headline , \" Donald Trump suggests President Obama was involved with Orlando shooting . \"\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"shooting\", \"life.injure.n/a\", \"An injury event involving human life, indicating harm or damage to a person's physical or mental state due to an incident.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nPentagon press secretary Peter Cook said in an emailed statement that Russian officials did not voice concerns earlier on Saturday when informed that coalition aircraft would be operating in the strike area .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"strike\", \"life.injure.n/a\", \"This event type indicates a potential injury or harm to a living being during a strike operation, though specific details are missing.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nPhotos : World War II in pictures U.S. troops assault Omaha Beach during the invasion of Normandy on June 6 , 1944 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"assault\", \"conflict.attack.n/a\", \"Represents a military attack during a conflict, involving forceful action against enemy forces in a combat scenario.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nDemocrats are big fans of same - day voting registration and early voting , and are thrilled with court decisions this past week they believe will help their turnout efforts this November .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"voting\", \"government.vote.castvote\", \"Voters cast their ballots on the same day they register or before election day, increasing accessibility and participation in democratic processes.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nRemember the loan you got , Ted . ’ \"\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"loan\", \"transaction.transfermoney.borrowlend\", \"Represents a financial borrowing event where one party receives funds with an obligation to repay.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMeanwhile , whistleblowing website WikiLeaks also said that it had acquired and would publish a copy of the \" cyber weapons \" cache shortly .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"acquired\", \"transaction.transferownership.n/a\", \"Represents a transfer of ownership where one party acquires control or rights over an asset or object.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBy the time she was honorably discharged , in 2009 , she had reached the rank of captain , according to military records .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"discharged\", \"personnel.endposition.firinglayoff\", \"This event denotes the official termination of a person's military service through discharge, typically with honors or under specific conditions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nPhotos : World War II in pictures Harry S. Truman takes the oath of office on April 12 , 1945 , as he becomes the 33rd president of the United States .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"oath\", \"contact.commitmentpromiseexpressintent.meet\", \"An official pledge or promise made by an individual to assume a leadership role, signifying formal commitment to office responsibilities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAnd yet here we are discussing a Republican candidate who insulted John McCain in front of the entire world -- claiming that John McCain is n't a hero because he was captured and spent time in a POW camp .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"captured\", \"movement.transportperson.preventexit\", \"An event where a person is prevented from exiting a transport, indicating restricted movement or forced detention.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nCountries with a high or moderate risk are being urged to take stronger measures to stop the spread of mosquitoes and reduce their numbers , including encouraging communities to get rid of pooled water that can form a breeding ground .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"urged\", \"contact.requestadvise.broadcast\", \"An official call to action urging countries to adopt stronger measures against mosquito breeding through community water management efforts.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nSaudi Arabia has perpetrated a mass execution that puts ISIS ’s beach beheadings to shame .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"execution\", \"justice.judicialconsequences.execute\", \"This event denotes a formal judicial killing of individuals, indicating state-sanctioned punishment or mass sentencing in a legal or political context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nJUST WATCHED Report : Colin Powell advised Clinton on email Replay More Videos ... MUST WATCH Report : Colin Powell advised Clinton on email 01:59\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"advised\", \"contact.requestadvise.meet\", \"Colin Powell sought advice from Clinton during a meeting regarding email-related matters.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ Al Qaeda has confiscated huge amounts of weapons from bases in Yemen , ” he said .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"confiscated\", \"transaction.transaction.n/a\", \"Represents the seizure or taking of assets or items from a location, indicating control or ownership transfer.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nDevin Stewart , a noted expert on Japan , and senior program director at the Carnegie Council for Ethics in International Affairs , suggests that an apology could “ open a can of worms on many , many issues . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"suggests\", \"contact.requestadvise.n/a\", \"An expert proposes a potential consequence or outcome of a proposed action or decision in a policy or diplomatic context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBut in practice , the government has taken back a number of areas with starve - or - surrender tactics , bombarding and starving people until they agree to leave .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"bombarding\", \"conflict.attack.airstrikemissilestrike\", \"A military airstrike or missile strike involving attacks on targets using aerial bombardment during conflict.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nA live audience of about 1,000 people gathered to watch the first Clinton - Trump debate , with a television audience predicted to reach as high as 100 million Americans .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"debate\", \"contact.discussion.meet\", \"A formal discussion between two candidates where they exchange views on policy and issues in a public setting.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nA great many will vote for the team they ’ve long been signed up with , while everybody else will be voting against one of these two extremely unattractive and unpopular candidates .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"voting\", \"government.vote.n/a\", \"A vote is being conducted by citizens in a political or governmental context, reflecting public opinion on candidates or policies.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe men put bags over her head and those of her husband and 17-year - old son , beating and kicking them while demanding repayment .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"kicking\", \"life.injure.injurycausedbyviolentevents\", \"An injury resulting from violent physical actions such as beating or kicking during confrontations or assaults.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHe took out full page ads asking for the death penalty to be brought back and applied to five innocent children and on Friday he doubled down on that .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"death penalty\", \"justice.judicialconsequences.execute\", \"This event represents the decision to carry out executions as a judicial consequence, reflecting a call to reinstate the death penalty.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAssange accused Clinton of trying to stir up ' neo - McCarthyist hysteria ' after the leaks showed the DNC had favored her over Bernie Sanders in the primary campaign\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"accused\", \"justice.initiatejudicialprocess.chargeindict\", \"This event represents the formal accusation of a criminal act initiated by a legal authority against an individual or entity.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHe added that his priority was to tackle drug trafficking and addiction - an ambitious plan in a country where a record number of people use methamphetamines , known in Thai as \" ya ba \" or \" crazy medicine \" , according to a 2015 U.N. report .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"trafficking\", \"movement.transportartifact.smuggleextract\", \"Involves illegal movement of drugs across borders or regions for distribution and consumption, often involving covert operations and networks.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nA few months later , in the November 2012 presidential election , Romney lost Florida by fewer than 75,000 votes , making consequential the 180,000 Hispanic voters who were on the voting roles despite serious questions about their eligibility to vote .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"vote\", \"government.vote.castvote\", \"A voter cast a ballot in a government election, indicating participation in a formal electoral process.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHide Caption 10 of 22 Photos : Istanbul airport attacked People stand outside the airport after the attack .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"attacked\", \"life.injure.injurycausedbyviolentevents\", \"This event captures injuries resulting from violent attacks, indicating harm caused by physical aggression in public or sensitive locations.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nOther ideas under consideration include sending more U.S. special operations forces to train and advise Kurdish and Syrian rebel groups , and deploying additional American and allied naval and airpower to the eastern Mediterranean , where a French aircraft carrier is already en route .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"advise\", \"contact.requestadvise.meet\", \"Requesting U.S. forces to train and advise Kurdish and Syrian rebel groups through direct contact and coordination.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAmerican and French soldiers were seen gathering on the scene before the counter - assault began , according to local reporters who tweeted out pictures of special forces by the hotel .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"assault\", \"conflict.attack.n/a\", \"An unclassified attack event indicating hostile military action initiated against a target, possibly involving combat forces in a conflict zone.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe FBI announced over the weekend that its investigation into a new batch of Hillary Clinton emails has found no criminal wrongdoing .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"investigation\", \"justice.investigate.investigatecrime\", \"An official probe is conducted to determine if criminal activity has occurred within a specific case or incident involving alleged illegal actions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHis comment was in reference to a New York Times story included in the email that described how “ The most generous donors to the Bill , Hillary & Chelsea Clinton Foundation will convene on June 6 at the Goldman Sachs . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"convene\", \"contact.discussion.meet\", \"A meeting between donors to the Clinton Foundation scheduled at Goldman Sachs on June 6 to discuss foundation activities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThis escalation had dire effects : The US now incarcerates more people than any other country in the world , including authoritarian regimes like China and Russia .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"incarcerates\", \"movement.transportperson.preventexit\", \"This event represents the prevention of a person's exit from a location, typically through restriction or enforcement.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n( Travis M. Andrews ) CDC officials are ramping up public awareness efforts on Legionnaires ’ disease , after cases of the easily - preventable sickness quadrupled in the U.S. over the past 15 years .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"sickness\", \"life.injure.n/a\", \"An injury event involving a person suffering harm from a disease, indicating a health-related incident with potential public health implications.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAs he has advocated from South Korea , Jung 's identity has been questioned by the North Korean government -- a common tactic used to target vocal defectors .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"advocated\", \"contact.requestadvise.broadcast\", \"An advocacy message broadcasted by an individual to influence public opinion or policy from a specific location.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nFranco Modigliani ’s presidential address to the American Economic Association in the late ' 70s was a remarkable white - flag - of - surrender speech by one of the prominent Keynesians .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"surrender\", \"conflict.yield.surrender\", \"A significant retreat or admission of defeat in a conflict, signaling a loss of position or resistance.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nOnce the layoffs start , workers will be willing to take big pay cuts to keep their jobs .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"layoffs\", \"personnel.endposition.firinglayoff\", \"Workers are terminated from their positions due to organizational downsizing or economic reduction in workforce.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMr. Trump also reiterated his view that NATO , the security alliance formed as a bulwark against the Soviet Union , was “ obsolete . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"reiterated\", \"contact.publicstatementinperson.n/a\", \"An in-person public declaration where an individual reiterates a previously expressed opinion or view to an audience or public.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAll along Mrs. Clinton publicly insisted she had no knowledge of the firings .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"firings\", \"personnel.endposition.firinglayoff\", \"An event indicating the termination of an employee's position through dismissal or layoff by an organization or individual.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nA damaged building is pictured after an airstrike on the rebel held al - Fardous neighbourhood of Aleppo , Syria February 18 , 2016 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"airstrike\", \"conflict.attack.airstrikemissilestrike\", \"An airstrike involves aerial bombardment using missiles or aircraft, targeting a location during conflict, causing damage to structures or people.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHer calm could well be because she has no doubt she 'll be found guilty and wants this stage of proceedings to be over .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"proceedings\", \"justice.initiatejudicialprocess.n/a\", \"The initiation of a judicial process begins formal legal proceedings against an individual or entity in a court of law.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBut Booth said an arms embargo could only work if neighboring countries cooperated .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"embargo\", \"transaction.transferownership.embargosanction\", \"An embargo sanction involves restricting trade or arms flow through cooperation among neighboring countries to enforce policy or security measures.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMany analysts who have watched Trump closely – and some who know him — think most of his seemingly self - destructive campaign behavior – for example his ongoing conflict with the family of a slain Muslim soldier , which has cost him his lead in the polls — is no accident .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"slain\", \"life.die.deathcausedbyviolentevents\", \"An individual dies due to violent actions, indicating a fatal outcome from conflict or aggression in the event context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nOf course , if Hillary Rodham Clinton would just hold a press conference , at which every question would feature those three words in some order or another , then we 'd all turn to discussing the comprehensive mental health plan that she released to thundering silence on Monday when most of the press was in an Anthony Weiner frenzy .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"discussing\", \"contact.discussion.correspondence\", \"An event where individuals engage in discussion or correspondence, exchanging ideas or information through verbal or written means.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ Right now there is a yawning gap in our lead - testing protocols , ” Sen. Charles E. Schumer ( D - N.Y. ) said in a statement last week when he introduced legislation that would provide $ 100 million in grants to help schools test drinking water for lead .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"grants\", \"transaction.transferownership.n/a\", \"This event represents a transfer of ownership or allocation of resources, such as grants, from one party to another in a formal or institutional setting.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIt was initially thought the judge had found her guilty , but her lawyers said a final verdict would be given on Tuesday .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"guilty\", \"justice.judicialconsequences.convict\", \"This event indicates a judicial finding of guilt against an individual in a legal proceeding, signifying legal responsibility for a crime.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nControversy 4 — Cover - up : Has Clinton misled investigations into her emails ?\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"misled\", \"contact.prevarication.correspondence\", \"This event involves intentional misleading of investigators through communications about email activities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn addition to renouncing his own tax plan , he said he ’s now open to raising the minimum wage and that he was no longer self - funding his campaign ( a point he constantly bragged about in the primary debates ) .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"funding\", \"transaction.transfermoney.giftgrantprovideaid\", \"This event represents a financial transfer involving aid, grants, or gifts provided to an individual or entity, often for campaign or policy support.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAnd though Obama shrugged off the comments from the \" colorful \" leader during his own news conference Monday , the White House later canceled his meeting with Duterte , conferring instead with the President of South Korea .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"conferring\", \"contact.discussion.meet\", \"A formal discussion or meeting between political leaders to exchange views or negotiate issues in a structured setting.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nPeople will know that you ca n't just smuggle in , hunker down and wait to be legalized .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"smuggle\", \"movement.transportperson.smuggleextract\", \"Involves illegally moving a person into a restricted area to evade legal processes or regulations.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHillary Clinton 's campaign Saturday compared the hacking into John Podesta 's emails and those from the Democratic National Committee as a new Watergate scandal while accusing GOP presidential nominee Donald Trump as being deeply involved in the plot , according to an essay posted on Medium Saturday .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"accusing\", \"justice.initiatejudicialprocess.chargeindict\", \"This event represents formally charging someone with a crime through judicial process, indicating legal action initiated against a person or entity.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAnd you will be deceived by Clinton .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceived\", \"contact.prevarication.broadcast\", \"An event where false information is spread to mislead or deceive individuals through public communication channels.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe new Clinton documents also include a censored email that shows Clinton and Cheryl Mills were contacted by Maria Haley , a longtime Clinton aide ( since deceased ) who was implicated the Chinagate scandal .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceased\", \"life.die.nonviolentdeath\", \"An individual died without violence, indicating a non-violent end to life, often due to natural causes or old age.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHide Caption 67 of 83 Photos : Inside North Korea Young children in a newly constructed Pyongyang orphanage practice a musical performance .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"constructed\", \"manufacture.artifact.build\", \"The creation of a physical structure or facility, such as a building or institution, through construction activities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ The notion that that anyone donating to the foundation was going to influence me in my job is absurd , ” Benenson suggested Clinton say , to which Margolis suggested , adding “ and never happened . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"donating\", \"transaction.transaction.giftgrantprovideaid\", \"A donation or financial contribution provided to an organization or individual for support or funding purposes.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nKerry said the Syrian regime had \" forced the surrender \" of Daraya in contravention of the February cessation of hostilities agreement , but Lavrov said the local accord was an \" example \" that should be \" replicated \" .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"surrender\", \"conflict.yield.surrender\", \"A military stronghold yields under pressure, indicating a loss of resistance in a conflict zone despite formal ceasefire agreements.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWith the planned study , The Daily Telegraph quotes from various doctors who have expressed serious doubts about the likelihood of Mr Spiridinov ’s brain remaining functional by the time the surgery is complete .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"surgery\", \"life.injure.n/a\", \"An injury to a person's body or mind during a medical procedure, potentially resulting in lasting harm or loss of function.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n‘ ‘ Hillary believes that it ’s vital to deceive the people by having one public policy and a totally different policy in private , ’ ’ he said to growing jeers .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceive\", \"contact.prevarication.meet\", \"An event where one party intentionally misleads another through conflicting public and private policies.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTo ensure that criminal aliens convicted of illegal reentry receive strong mandatory minimum sentences .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"convicted\", \"justice.judicialconsequences.n/a\", \"This event indicates a legal finding of guilt against an individual for violating immigration laws through illegal reentry.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWe can keep going for two to three months , but then people will starve in large numbers .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"starve\", \"life.injure.n/a\", \"This event represents a life-threatening injury or harm resulting in physical or survival-related damage to individuals over time.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBut in practice , the government has taken back a number of areas with starve - or - surrender tactics , bombarding and starving people until they agree to leave .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"surrender\", \"conflict.yield.surrender\", \"A conflict resolution where one side yields due to pressure, often through coercion, leading to surrender without resistance or fight.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAn inspection team led by Independent Chair Rune Andersen , an independent international anti - doping expert ( Norwegian ) and three members of the IAAF Council who will be appointed in the next few days.’\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"inspection\", \"inspection.sensoryobserve.inspectpeopleorganization\", \"An inspection team observes people and organizational structures to assess compliance with anti-doping regulations and standards.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe Amaq news agency , which is linked to Islamic State , said IS militants had detonated a car bomb and then blown up explosive belts .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"detonated\", \"conflict.attack.n/a\", \"Represents an attack carried out by militants involving explosives, indicating violent action against a target or group in a conflict zone.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHe and others also said a Trump win in November wo n't necessarily repel future recruits nor lead to mass retirements , at least not immediately .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"recruits\", \"personnel.startposition.hiring\", \"This event signifies the beginning of a new role or position through recruitment, indicating the addition of personnel to an organization.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBryant is suspected of vandalizing signs at the property several other times in previous months .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"vandalizing\", \"artifactexistence.damagedestroy.n/a\", \"This event indicates intentional damage to physical objects, such as signs, through malicious acts like defacement or destruction.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nYet according to a financial statement filed as evidence in her case , Palmer started working in the United States nine months before this authorization was granted .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"granted\", \"transaction.transaction.giftgrantprovideaid\", \"This event represents the act of granting aid or resources, typically involving a transfer of value or support to an individual or entity.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIt is why we call for the support and strengthening of the BDS movement and for the boycott of Israeli goods .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"boycott\", \"transaction.transferownership.embargosanction\", \"This event represents the imposition of economic sanctions through boycotts targeting Israeli goods to support the BDS movement.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe claims by Hezbollah sources come even as the Gulf Cooperation Council ( GCC ) on Wednesday declared the group a terrorist organization , citing Hezbollah 's activity in the Gulf states recruiting terrorists and smuggling weapons - a point that was not raised in the article .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"smuggling\", \"movement.transportartifact.smuggleextract\", \"Involves the illegal transfer of weapons or contraband across borders, often to support militant groups or destabilize regions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nScholars estimate that under Stalin , more than 1 million people were executed in political purges and millions more died as a result of harsh labor and cruel treatment in the vast gulag prison camp system , mass starvation in Ukraine and southern Russia , and deportations of ethnic minorities .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"starvation\", \"life.injure.illnessdegradationhungerthirst\", \"This event represents physical harm due to lack of food, leading to illness, weakness, or death in individuals under oppressive conditions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMoreover , finding millions of undocumented migrants almost certainly would entail immigration agents entering homes , raiding businesses and operating roadblocks to check identity papers to separate U.S. citizens and approved immigrants from those in the country illegally , a winnowing - out process that undoubtedly would be challenged in court .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"roadblocks\", \"movement.transportartifact.prevententry\", \"Preventing entry by blocking transportation routes to stop unauthorized movement into a jurisdiction.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThis escalation had dire effects : The US now incarcerates more people than any other country in the world , including authoritarian regimes like China and Russia .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"incarcerates\", \"justice.arrestjaildetain.arrestjaildetain\", \"The event represents the act of capturing and holding individuals in prison, indicating a systemic pattern of incarceration within a nation's justice system.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nPartially , it is a symptom of the superior public relations enjoyed by Silicon Valley technology corporations , and the fact that those corporations also provide the bulk of private funding for the flagship digital privacy advocacy groups , leading to a conflict of interest .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"funding\", \"transaction.transaction.n/a\", \"Represents financial support provided by corporations to advocacy groups, indicating a funding relationship with potential conflict of interest.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe battle lasted more than five months , ending in February 1943 , at the cost of at least 160,000 German soldiers killed or captured .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"captured\", \"movement.transportperson.preventexit\", \"This event describes preventing a person from exiting a transport, typically during military or forced movement scenarios.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nPresident Salva Kiir 's sacking of Machar as his deputy ignited a war in December 2013 which has killed thousands and displaced millions in the world 's newest country .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"sacking\", \"personnel.endposition.firinglayoff\", \"Removal of a person from a leadership or official position, typically through dismissal or termination by authority.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThey also said there should be an \" appropriate dialogue with Russia to be maintained to enhance transparency and reduce risk . \"\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"dialogue\", \"contact.discussion.correspondence\", \"An exchange of communication between parties to discuss issues, promote transparency, and manage risks through structured dialogue.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIt remains a mystery why an FBI public corruption investigation into Mrs. Clinton ’s tenure as Secretary of State has not been front - page news since it was first revealed in January .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"investigation\", \"justice.investigate.investigatecrime\", \"An official probe into illegal activities involving a public official during their government service, indicating potential wrongdoing.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nI think offering people some form of early voting makes sense .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"voting\", \"government.vote.castvote\", \"This event represents individuals casting votes in a governmental election process through early voting arrangements.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ Today we solemnly reflect on the first mass atrocity of the 20th century — the Armenian Meds Yeghern — when one and a half million Armenian people were deported , massacred , and marched to their deaths , ” he said in a statement .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"massacred\", \"life.die.deathcausedbyviolentevents\", \"This event represents the death of individuals due to violent acts, such as massacres, during historical atrocities involving mass killing and forced displacement.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIt calls on the WHO to convene an independent group to advise it and the International Olympic Committee , and for authorities to reconsider the decision to hold the Games in Rio .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"convene\", \"contact.discussion.correspondence\", \"An official communication request for a group to be formed to provide advice to organizations involved in global health and sports governance.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAfter the embargo began in 2014 , Valio sold its \" Oltermanni \" cheese , popular in Russia , for half its regular price in Finland because of oversupply .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"embargo\", \"transaction.transaction.embargosanction\", \"An embargo imposition restricts trade or exports, often due to political or economic reasons, affecting commercial transactions between nations.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWhen a family grows concerned about its future income or job prospects , it can not stop paying rent or other necessities , but it typically defers spending on durable goods such as cars , appliances , luxury items and vacations .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"rent\", \"transaction.transferownership.borrowlend\", \"This event represents a financial transfer where one party borrows or lends ownership of assets, typically involving rent or property.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nObama ’s refusal to call the mass killings a genocide is a gesture to Turkey , a NATO ally of the U.S. and an important member of the president ’s coalition fighting the Islamic State in Iraq and Syria ( ISIS ) .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"genocide\", \"life.die.deathcausedbyviolentevents\", \"This event represents death resulting from violent or mass violent actions, such as massacres or genocides, involving large-scale harm.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nShe added that she and French President Francois Hollande were working \" with all one 's strength \" to urge Ukraine and Russia to implement the so - called Minsk ceasefire agreement despite all the difficulties .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"urge\", \"contact.commandorder.broadcast\", \"An event where individuals issue strong calls to action to influence others toward a specific course of behavior or agreement.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn February , Assange , who is living in the Ecuadorian Embassy in London and is avoiding a rape charge in Sweden he claims is politically motivated , published an editorial on WikiLeaks calling Clinton a “ war hawk with bad judgment who gets an unseemly emotional rush out of killing people . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"rape\", \"life.injure.injurycausedbyviolentevents\", \"An injury resulting from violent events involving physical harm or aggression, often linked to criminal or conflict-related actions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIt is also approximately half the homicide rate in the United States .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"homicide\", \"life.die.deathcausedbyviolentevents\", \"This event represents a death resulting from violent acts, specifically homicides, involving human fatalities due to intentional killing.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTrump made waves in March during a GOP presidential debate when he called the 1989 Tiananmen Square Massacre , in which hundreds of protesters were killed as the People 's Liberation Army enforced martial law , a \" riot . \"\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"Massacre\", \"life.die.deathcausedbyviolentevents\", \"This event represents deaths resulting from violent conflicts or massacres, indicating loss of life due to armed confrontation or state-enforced violence.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn a recent development , Iran 's Foreign Ministry has urged Azerbaijan and Armenia to show restraint and avoid any move that can escalate their tensions .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"urged\", \"contact.requestadvise.broadcast\", \"A formal appeal by a state ministry to other nations to refrain from actions that could worsen regional tensions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHopes flickered among some anti - Trump Republicans that there would be a revolt against him when delegates convene to nominate him formally in Cleveland from July 18 to 21 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"convene\", \"contact.discussion.meet\", \"A meeting or discussion occurs between individuals to exchange views or plan actions, indicating interaction and potential alignment on political or strategic matters.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nRep. Mark Sanford ( R - S.C. ) , a professed Trump backer ( for now ) , wrote a Sunday New York Times op - ed titled , “ I Support You , Donald Trump .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"professed\", \"contact.publicstatementinperson.n/a\", \"A public in-person statement made by a political figure expressing support or opinion in a formal setting.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nLobbyist obstruction is largely the reason why we have n't studied the full effects of GMO crops or how they affect the livestock that eat them .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"obstruction\", \"movement.transportperson.prevententry\", \"Preventing a person from entering a location through obstruction, typically involving physical or bureaucratic barriers.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn 2011 , a major study by University of California , Berkeley , political scientist M. Steven Fish presented cross - national statistical data showing that between 1994 and 2007 , annual homicide rates in the Muslim world averaged just 2.4 per 100,000 of the population .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"homicide\", \"life.die.deathcausedbyviolentevents\", \"This event represents a death resulting from violent acts, specifically homicides, in a given population over a defined time period.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBut the pilferage of old yoga schedules and plans for a wedding long - consummated hardly seem the kind of thing to require the involvement of an old CIA director long - retired .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"retired\", \"personnel.endposition.n/a\", \"Denotes the cessation of a person's official position or role within an organization, indicating retirement from duty.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTwo of the former Trump models said Trump ’s agency encouraged them to deceive customs officials about why they were visiting the United States and told them to lie on customs forms about where they intended to live .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceive\", \"contact.prevarication.broadcast\", \"An instance where individuals are encouraged to misrepresent information to authorities during official interactions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe four - day battle became a major victory for the U.S. Navy , which sunk four Japanese aircraft carriers , and it marked a turning point in the war in the Pacific .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"battle\", \"conflict.attack.n/a\", \"A military engagement where one side attacks another, resulting in damage or loss, typically involving combatants and strategic outcomes.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nDavutoğlu condemned the attacks in Aleppo as “ barbarity , tyranny , a war strategy conducted with a medieval mentality ” and said hundreds of thousands of people in the region faced the danger of starvation if a humanitarian corridor was not opened .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"starvation\", \"life.injure.illnessdegradationhungerthirst\", \"This event represents severe physical harm due to lack of food, leading to degradation of health and potential death in a vulnerable population.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nDeadly force : The Russian diplomat was in this X6 SUV when a robbery attempt turned into a wrestling match which ended in the death of the robber\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"robbery\", \"conflict.attack.stealrobhijack\", \"A robbery attempt escalated into a violent confrontation resulting in physical harm or death to the attacker during a theft incident.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe New York Times reported on Sunday that a Ukrainian anti - corruption agency found Manafort ’s name on a secret ledger used by the ousted , pro - Russia government of ex - Ukrainian President Viktor Yanukovych .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"ousted\", \"personnel.endposition.firinglayoff\", \"This event signifies the removal of a person from their official position, typically due to dismissal or termination by an organization or government.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nOn Sunday , Trump noted the death of Elie Wiesel , a Holocaust survivor and author who fought for peace , human rights : \" On Saturday a great man , Elie Wiesel , passed away .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"passed away\", \"life.die.nonviolentdeath\", \"This event represents a non-violent death of a notable individual, emphasizing peaceful passing rather than violence or conflict.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nLearning that pollution produced by corporate greed was the major “ underlying driver ” behind many of the expanding illnesses ( especially asthma , diabetes , cancer , and learning disabilities ) she was confronting on an individual basis , she became a leading environmental and public health policy expert and advocate .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"illnesses\", \"life.injure.n/a\", \"An individual suffers physical or health-related harm due to external factors, indicating a negative health outcome from environmental or social causes.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAs for his supporters ’ tendency to sucker - punch people of color at rallies , Trump ’s supporters are happy to paint him as the victim , as Media Matters has shown ; for example , conservative radio host Wayne Allyn Root argued Monday that “ it ’s not [ Trump ’s ] fault that he ’s being victimized at rallies . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"punch\", \"life.injure.injurycausedbyviolentevents\", \"This event represents physical harm caused by violent actions during public gatherings, specifically involving punches directed at individuals of color.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHillary Clinton , for instance , talks constantly about her fears that families will be separated , but she 's not talking about the American families who have been permanently separated from their loved ones because of a preventable homicide , because of a preventable death , because of murder .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"homicide\", \"life.die.deathcausedbyviolentevents\", \"A person dies due to a violent act such as homicide, resulting in a preventable death caused by intentional violence.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nI ’m quite certain that most of the following of Donald Trump , for example , would want him in prison for life at best and hanged at worst .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"hanged\", \"conflict.attack.n/a\", \"Represents a violent conflict where one party inflicts harm on another, often involving extreme punishment or retaliation.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n\" The wholesale slaughter of families of terrorists — it is understood as a war crime , \" an Army intelligence officer told POLITICO .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"slaughter\", \"life.die.deathcausedbyviolentevents\", \"Refers to the death of individuals due to violent actions, particularly in contexts involving warfare or conflict.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBy the time he was done , he had misled large numbers of people into believing that he found proof that : 1 ) the documents in the WikiLeaks archive were altered ; 2 ) Russia put forgeries into the WikiLeaks archive ; 3 ) Sputnik knew about the WikiLeaks archive ahead of time , before it was posted online ; 4 ) WikiLeaks coordinated the release of the documents with the Russian government ; and 5 ) the Russian government and the Trump campaign coordinated to falsely attribute Eichenwald ’s words to Blumenthal .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"misled\", \"contact.prevarication.meet\", \"An event where an individual intentionally misled others through false information during a contact or interaction.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWASHINGTON ( Reuters ) - The United States is prepared to consider imposing sanctions or an arms embargo on South Sudan 's leaders if they fail to cooperate in a proposed unity government to end conflict in the country , a senior U.S. official said on Wednesday .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"embargo\", \"transaction.transferownership.embargosanction\", \"The U.S. may impose trade restrictions on South Sudan's leaders if they fail to cooperate in a unity government to end conflict.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTwo of the former Trump models said Trump ’s agency encouraged them to deceive customs officials about why they were visiting the United States and told them to lie on customs forms about where they intended to live .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceive\", \"contact.prevarication.correspondence\", \"Involves intentional misrepresentation in communication to mislead customs officials about visitation or residency intentions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBut we now have an obligation to them and to their children to control future immigration as we are following , if you think , previous immigration waves .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"immigration\", \"inspection.sensoryobserve.inspectpeopleorganization\", \"Observing individuals or organizations during immigration processes to assess compliance and manage future immigration flows.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn announcing that he would recommend against prosecution in July , Comey made a point of declaring that “ this is not to suggest that in similar circumstances , a person who engaged in this activity would face no consequences .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"recommend\", \"contact.requestadvise.broadcast\", \"An official communication where a person requests advice or makes a recommendation on a legal or policy matter to an audience or authority.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBut these charges — including the pricey rent that Blais and her roommates paid — consumed nearly all her modeling earnings .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"rent\", \"transaction.transferownership.borrowlend\", \"Represents a financial transaction where one party lends or borrows assets, such as rent payments between tenants and landlords.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAn Armenian soldier killed and allegedly decapitated by Azerbaijani troops in Nagorno - Karabakh has been reburied after the Armenian military has retrieved his severed head and handed it to his family .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"decapitated\", \"life.die.deathcausedbyviolentevents\", \"An individual dies due to violent actions by another party, indicating a fatal outcome from conflict or aggression in a specific location and time.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAnd though Obama shrugged off the comments from the \" colorful \" leader during his own news conference Monday , the White House later canceled his meeting with Duterte , conferring instead with the President of South Korea .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"conferring\", \"contact.discussion.correspondence\", \"An official exchange of views or discussion between political leaders through formal communication or meeting arrangements.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBut Trump has n’t backed off , repeating time and again his admiration for Russian President Vladimir Putin and his leadership style , and suggesting the United States work more closely with Moscow .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"suggesting\", \"contact.requestadvise.broadcast\", \"An event where one party proposes a new relationship or approach, suggesting collaboration or advice to another entity.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nRussia seemingly has ignored Secretary of State John Kerry ’s appeals to stop bombing civilians and allow critical humanitarian aid to starving Syrians – and is instead escalating its military involvement , deploying four of its most capable fighter jets to Syria , two defense officials confirmed to Fox News .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"starving\", \"life.injure.illnessdegradationhungerthirst\", \"This event represents severe physical suffering due to lack of food and water, leading to degradation of health and life.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe ICC , which began hearing cases at The Hague in the Netherlands in 2002 , was established as a court of last resort , to prosecute grave abuses when countries would not or could not act against those responsible .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"established\", \"government.formation.n/a\", \"The establishment of a governmental body as a final judicial authority to handle serious international crimes when national systems fail.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe batch of emails — which Wikileaks promised is the first of many more to come — provided a glimpse into the inner workings of the campaign , and offered telling details about Clinton ’s views on trade and the middle class .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"promised\", \"contact.commitmentpromiseexpressintent.meet\", \"An entity expresses intent to commit or meet in the future, indicating a promised action or engagement.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nGuccifer 2.0 had always claimed to be Romanian like the original Guccifer , imprisoned hacker Marcel Lazăr Lehel , but the 2.0 version had regularly communicated with journalists in Russian .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"imprisoned\", \"justice.arrestjaildetain.arrestjaildetain\", \"This event signifies the detention or imprisonment of an individual by legal authorities, indicating a formal legal action taken against a person.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe new Clinton documents also include a censored email that shows Clinton and Cheryl Mills were contacted by Maria Haley , a longtime Clinton aide ( since deceased ) who was implicated the Chinagate scandal .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceased\", \"life.die.nonviolentdeath\", \"An individual died without violence, indicating a non-violent end to life, often due to natural causes or old age.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIt 's an incident \" likely to result in demonstrable harm to the national security interests , foreign relations , or economy of the United States or to the public confidence , civil liberties , or public health and safety of the American people . \"\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"harm\", \"life.injure.n/a\", \"An event involving physical or psychological harm to individuals, potentially affecting national security, public health, or civil liberties.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBoth Rubio and Kasich appear now to be angling for a seat at a contested convention , in which they block Trump from getting 1,237 , and potentially emerge victorious by negotiating at the gathering in Cleveland in July .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"negotiating\", \"contact.discussion.correspondence\", \"This event represents a formal or informal discussion between political figures to reach an agreement on convention strategy and positioning.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nSunni tribal leaders , who had always hated living under AQI 's harsh and often violent rule , became convinced that the Shias were starting to win Iraq 's sectarian civil war .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"violent\", \"life.injure.injurycausedbyviolentevents\", \"An injury resulting from violent actions during conflict, indicating physical harm caused by aggressive or hostile events in a war context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nEpstein pleaded guilty to soliciting a minor for prostitution in 2008 , and served just 18 months in jail for the offense despite outrage from victims advocates .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"guilty\", \"justice.judicialconsequences.convict\", \"An individual is found legally responsible for a crime through a guilty plea, resulting in judicial consequences and legal accountability.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAt State , on the Policy Planning Staff , Cohen was soon christened “ Condi ’s party - starter , ” channeling buzzwords from Silicon Valley into US policy circles and producing delightful rhetorical concoctions such as “ Public Diplomacy 2.0 . ” On his Council on Foreign Relations adjunct staff page he listed his expertise as “ terrorism ; radicalization ; impact of connection technologies on 21st century statecraft ; Iran . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"producing\", \"manufacture.artifact.n/a\", \"The creation of a new rhetorical or conceptual product influenced by technological or digital culture in policy discourse.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nUday reputedly learned the trade of terror from Saddam , who took him as a child to tour the prisons where his regime tortured and executed hundred of thousands of Iraqis .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"executed\", \"life.die.deathcausedbyviolentevents\", \"Death resulting from violent actions such as torture or execution by a regime, indicating intentional and harmful state violence.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIf you live in such a country , obtain an assault rifle legally , preferably AK-47 or variations . \"\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"obtain\", \"transaction.transferownership.n/a\", \"This event represents the legal transfer of ownership of an assault rifle, such as an AK-47, from one party to another.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHis comment was in reference to a New York Times story included in the email that described how “ The most generous donors to the Bill , Hillary & Chelsea Clinton Foundation will convene on June 6 at the Goldman Sachs . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"convene\", \"contact.discussion.correspondence\", \"An exchange of communication involving discussion or contact between individuals regarding a shared topic or event.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nProud of her role in superintending the sanctions regime that led to the death of some half million Iraqis , Albright deserves a special place in hell too .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"sanctions\", \"transaction.transferownership.embargosanction\", \"This event represents the transfer of ownership or control over sanctions imposed on a country, restricting its trade or activities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nSeven Christian homes were also ransacked and torched in the incident .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"ransacked\", \"artifactexistence.damagedestroy.n/a\", \"This event indicates the unlawful entry and destruction of property, involving forced removal and damage to homes.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nSecretary of State John Kerry said Syrian forces hit a hospital overnight , killing 20 people and wounding 100 , describing what would be the latest strike by Russia or its ally in Damascus on a civilian target .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"strike\", \"life.injure.injurycausedbyviolentevents\", \"Violent events caused injuries to civilians, including deaths and wounds, resulting from attacks on medical facilities during conflict.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe EU 's leaders in Brussels are expected to play hardball in negotiating Britain 's exit , to send a message to other states that might be contemplating a similar move .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"negotiating\", \"contact.discussion.meet\", \"Leaders engage in direct talks to negotiate complex political issues, such as Brexit, to influence outcomes and demonstrate firmness.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWhat Donald Trump is saying that he would unilaterally surrender to Russia and Putin , give Putin a massive foreign policy victory by breaking NATO and abandoning Europe . \"\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"surrender\", \"conflict.yield.n/a\", \"An event where a party gives up a strategic position, leading to a favorable outcome for an opponent in a geopolitical conflict.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIf crime rates remained the same ( which , of course , they did n’t ) , 83 percent of all Americans aged 12 at the time would , in their actuarial lifetimes , be victimized by an attempted or completed violent crime , and 40 percent would be injured as a result of a robbery or assault .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"robbery\", \"conflict.attack.n/a\", \"This event represents a violent attack incident involving physical harm or attempted harm to individuals in a societal context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMore than 100 prominent physicians , bioethicists and scientists from around the world posted a letter Friday urging World Health Organization Director - General Margaret Chan to exert pressure on Olympic authorities to move the Olympics from Rio de Janeiro or delay the Games because of public health concerns over the Zika virus .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"urging\", \"contact.requestadvise.correspondence\", \"A formal request for advice or action through correspondence, expressing concern and seeking influence on a decision.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe United States - led bombing campaign against the Islamic State in Iraq and Syria surpassed a stunning milestone this month , underscoring the magnitude of the coalition ’s quest to crush the militant group .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"bombing\", \"conflict.attack.bombing\", \"A military bombing campaign targeting a militant group in a conflict zone to weaken or destroy the organization through aerial attacks.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn other words , Amiri ’s assistance to the US had been reported publicly for six years before his execution , years before the release of any emails from state department servers .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"execution\", \"justice.judicialconsequences.execute\", \"The event represents the judicial execution of an individual, indicating a formal and final punishment by state authority.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn 2015 , another front opened up in Yemen , where the Saudis began a US - supplied brutal air war against that country ’s new Shiite Houthi government and a starvation blockade of its people .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"starvation\", \"life.injure.illnessdegradationhungerthirst\", \"An event where individuals suffer physical or mental harm due to lack of food, leading to degradation, illness, or death in a population.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nOnce the layoffs start , workers will be willing to take big pay cuts to keep their jobs .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"layoffs\", \"personnel.endposition.n/a\", \"This event signifies the termination of employment positions within an organization, leading to job losses for workers.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHow can he be guilty as an agent of an enemy power , when his goal was to defend the people of the United States against secret and illegal actions by their own government ?\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"guilty\", \"justice.judicialconsequences.convict\", \"This event indicates a legal finding of guilt against an individual for actions deemed illegal or harmful by the justice system.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHe was murdered by an illegal in 2010 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"murdered\", \"life.die.deathcausedbyviolentevents\", \"This event represents a death caused by violent actions, specifically murder, indicating intentional and fatal harm to an individual.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTrump also faces questions about what happened to money from his most recent book , “ Crippled America . ” He pledged to donate the profits to charity : “ In May , Trump reported earning more than $ 1 million in royalties from the book .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"donate\", \"transaction.transfermoney.giftgrantprovideaid\", \"An act of giving money to another party for charitable or supportive purposes without expectation of return.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWhile no American can board an airplane , rent a car or even cash a check without showing a valid government - issued ID , the Democratic Party charges “ voter suppression ” and brings lawyers to court to litigate every time efforts are made to pass basic voter ID laws .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"rent\", \"transaction.transferownership.borrowlend\", \"This event represents a transfer of ownership or borrowing arrangement involving a tangible item like a car through rental.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHead transplantation involves decapitating the patient . Pushing forward his idea , Canavero has now proposed a two - part procedure .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"decapitating\", \"life.die.deathcausedbyviolentevents\", \"This event represents a violent death caused by decapitation, indicating a severe and intentional act leading to the patient's demise.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThere are only 35 doctors left in the district caring for hundreds of wounded , the World Health Organisation ( WHO ) said on Tuesday .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"wounded\", \"life.injure.injurycausedbyviolentevents\", \"This event represents injuries caused by violent events, indicating harm resulting from physical or violent actions in a community or region.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nEcuador on Tuesday vowed to remain mostly hands - off on the divisive issue , reaffirming the country 's decision to grant Assange asylum “ based on his legitimate fears of persecution because of his journalistic activities as the editor of WikiLeaks . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"asylum\", \"movement.transportperson.grantentryasylum\", \"Granting asylum involves allowing a person to enter and stay in a country due to fear of persecution based on their activities or beliefs.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBut three days after that call , the FBI Cyber Division issued a potentially more disturbing warning , titled “ Targeting Activity Against State Board of Election Systems . ” The alert , labeled as restricted for “ NEED TO KNOW recipients , ” disclosed that the bureau was investigating cyberintrusions against two state election websites this summer , including one that resulted in the “ exfiltration , ” or theft , of voter registration data .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"investigating\", \"justice.investigate.investigatecrime\", \"The FBI is conducting a formal investigation into cyberattacks targeting state election systems for possible criminal activity.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nI think we would have forced them to the negotiating table .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"negotiating\", \"contact.negotiate.correspondence\", \"This event represents formal communication between parties to reach mutual agreement through structured dialogue or correspondence.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nNeela Ghoshal , a Nairobi - based activist for Human Rights Watch , told The Daily Beast that “ it ’s hard to imagine how a practice that has been harshly condemned by the UN Special Rapporteur on Torture can continue to be authorized in Kenya . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"Torture\", \"life.injure.injurycausedbyviolentevents\", \"This event represents physical harm caused by violent acts, specifically torture, as recognized by international human rights bodies.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBERLIN There is no reason to lift the European Union 's sanctions against Russia as Moscow has not fulfilled all of its commitments under an international peace plan , German Chancellor Angela Merkel said in an interview published on Friday .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"sanctions\", \"transaction.transaction.embargosanction\", \"Imposing restrictions on trade or activities against a country due to violations of international agreements or commitments.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe men put bags over her head and those of her husband and 17-year - old son , beating and kicking them while demanding repayment .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"demanding\", \"contact.commandorder.broadcast\", \"An event where an agent issues a command or demand to a victim, often in a coercive or violent context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ According to them , vacuum - based massage improves circulation and overall well being , suggesting that muscle repair happens faster after physical exertion .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"suggesting\", \"contact.requestadvise.correspondence\", \"This event represents a suggestion made through correspondence about the benefits of vacuum-based massage on muscle repair and well-being.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHowever , it was Mutko 's deputy , Yuri Nagornykh , who was advised of \" every positive analytical finding \" from the Moscow laboratory from 2011 onwards - in \" total violation \" of Wada rules - and decided who to protect .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"advised\", \"contact.requestadvise.n/a\", \"An official received analytical findings from a lab and made decisions on who to protect, violating Wada rules.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nCNAS is the most influential foreign policy think tank in Washington and is advised , led and staffed by almost all of the key players in Washington ’s foreign policy and military establishment — from both parties .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"advised\", \"contact.requestadvise.correspondence\", \"An entity seeks advisory input from another through formal or informal correspondence in foreign policy and military contexts.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThose concerns prompted Homeland Security Secretary Jeh Johnson to convene a conference call with state election officials on Aug. 15 , in which he offered his department ’s help to make state voting systems more secure , including providing federal cybersecurity experts to scan for vulnerabilities , according to a “ readout ” of the call released by the department .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"convene\", \"contact.discussion.meet\", \"An official meeting initiated to discuss issues related to election security and coordination among state officials.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWe will terminate the Obama administration 's deadly , and it is deadly , non - enforcement policies that allow thousands of criminal aliens to freely roam our streets , walk around , do whatever they want to do , crime all over the place .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"walk\", \"movement.transportperson.selfmotion\", \"An individual moves independently through space without external transportation, indicating personal locomotion in a physical environment.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHe said Russia was “ intent on allowing the killing to continue and , indeed , participating in carrying it out ” and that what was needed from Moscow was “ less talk and more action from them to stop the slaughter . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"slaughter\", \"life.die.deathcausedbyviolentevents\", \"Death resulting from violent acts or conflicts, indicating loss of life due to aggression or warfare.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHe noted that the political world , “ which today seems to lack a response in face of the massacre of Christians , will also have to fight ISIS and it will do it in a different way .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"massacre\", \"life.die.deathcausedbyviolentevents\", \"Death resulting from violent acts such as massacres, indicating a fatal outcome due to intentional or violent conflict.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe MoD further warned that the Obama regime was planning to kill President Erdogan in a coup and replace him the CIA ’s “ designated figurehead ” Fethullah Gulen — and that Turkey ’s MoFA confirmed just hours ago was behind this plot by stating that “ the Fethullah Gülen Terrorist Organization is behind this coup attempt ” .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"coup\", \"conflict.attack.n/a\", \"A sudden, violent takeover of government by force, often involving overthrowing a leader or regime through military or political action.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHillary Clinton , Stein agrees with me , has considerably less capacity to deceive and bamboozle progressive and young voters than Barack Obama enjoyed in 2007 - 08 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceive\", \"contact.prevarication.meet\", \"An event where individuals engage in deceptive communication, misleading others through false or hidden information during interaction.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n\" It is impossible for the FBI not to recommend criminal charges against Hillary Clinton , \" Trump tweeted Saturday .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"recommend\", \"contact.requestadvise.broadcast\", \"An official recommendation is publicly communicated to advise action, such as recommending criminal charges in a political context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nFrance ’s highest court of criminal appeals upheld the conviction of a dozen Palestine solidarity activists for publicly calling for the boycott of Israeli goods .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"boycott\", \"transaction.transferownership.embargosanction\", \"This event represents a formal restriction on trade or goods, enforced through legal or policy measures targeting specific entities or countries.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nU.S. authorities charged 46 people , ranging in age from 24 to 80 , with extortion , loan sharking , smuggling , arson , gun trafficking and other crimes in a federal indictment unsealed in New York .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"arson\", \"conflict.attack.setfire\", \"Involves intentional setting of fire to property as an act of violence or destruction in a conflict context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHillary Clinton ’s legacy in Iraq , Libya , and Syria has produced only turmoil and suffering , ” he said in Pennsylvania , skipping over his own one - time support for intervention in some of those conflicts .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"conflicts\", \"conflict.attack.n/a\", \"Represents instances of violence or warfare occurring in specific regions, indicating active conflict and potential harm to civilians or governments.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe Vladimir Putin satrap leveraged his Donbas mobster background to defeat , with Kremlin backing , the pro - democracy Orange Revolution and went on to build a kleptocracy underpinned by Russia ’s security services , which is believed to have trained the Ukrainian snipers who slaughtered more than 50 Maidan protesters in 2014 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"slaughtered\", \"life.die.deathcausedbyviolentevents\", \"Death resulting from violent attacks, indicating a lethal event involving physical harm and aggression in a specific context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThere is a tension between building up support for action against the Iranian nuclear program now and delivering the kind of reassurance that will be necessary once the capability has been acquired .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"acquired\", \"transaction.transferownership.n/a\", \"Represents a shift in control or possession of an asset, indicating a change in ownership rights or stewardship of a given entity.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWe 'll fight all who aim to divide Syria or harm its people .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"harm\", \"life.injure.injurycausedbyviolentevents\", \"This event represents physical or psychological harm caused by violent actions targeting individuals or groups in a specific location or time.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ We thought about how to improve [ the work of law enforcement ] in all areas , including those related to fighting terrorism , to organized crime and illicit drug trafficking , ” Putin said .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"trafficking\", \"movement.transportartifact.smuggleextract\", \"Involves illegal movement of goods or substances across borders for illicit profit, often involving organized networks and enforcement challenges.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn his ABC interview , Trump dismissed those Democratic claims .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"interview\", \"contact.discussion.meet\", \"An interaction between individuals involving conversation or discussion, typically in a formal or public setting.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAnd so I think an organized effort at voter intimidation coming from the Trump campaign seems very unlikely .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"voter intimidation\", \"government.vote.violationspreventvote\", \"This event denotes illegal actions that prevent individuals from voting by threatening or coercing them through intimidation.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nClinton was walking away from the lectern after a press conference when a reporter yelled a question about her GOP opponent granting an interview to a Russian - funded news network .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"granting\", \"transaction.transferownership.n/a\", \"Denotes the act of transferring control or rights to an entity, such as authorizing access or permission in a transaction.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn fact , the reverse is the case : The higher the percentage of Muslims in a society , the lower the homicide rate .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"homicide\", \"life.die.deathcausedbyviolentevents\", \"This event represents a death resulting from violent acts, specifically homicides, within a societal context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n1:29 Budget bills ' rushed through in the dark of night , ' GOP complains 0:50 ​Lawmaker demonstrates how to rescue a hot dog 1:02 Firing Trump - Russia special counsel ' would be too reminiscent of Watergate ' 1:19 Don Bachardy tells the story behind his Capitol portrait of Jerry Brown 1:28 Josh Newman vows to beat recall attempt : ' We are not giving this seat back ' 1:01 ' There 's really no check ' on the populous parts of California , assemblyman says 1:23 Democrats need to win back hearts and minds , Kimberly Ellis tells activists 1:11 California senator argues for single - payer health care in California Share Video Video link : Select Embed code : Select\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"Firing\", \"personnel.endposition.firinglayoff\", \"This event represents the termination of a person's position, typically due to organizational or political decisions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nNigerian - born doctor Bennet Omalu , who has worked extensively on concussions , suggested that Clinton had been poisoned by Putin and Trump .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"suggested\", \"contact.requestadvise.correspondence\", \"An individual proposes a claim or opinion regarding a situation through communication or correspondence with another party.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWe now know breakout time would be reduced to zero in as little as 10 years and there are not true anytime / anywhere inspections .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"inspections\", \"inspection.sensoryobserve.inspectpeopleorganization\", \"This event involves sensory observation of people or organizations during inspections, assessing their conditions or activities through direct perception.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMoments later , Mr. Trump stood by his past remarks that he would have supported the impeachment of President George W. Bush over the war in Iraq , saying that his administration misled the country about weapons of mass destruction .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"misled\", \"contact.prevarication.broadcast\", \"An official spreads false information to mislead the public about government decisions or actions, particularly regarding war or policy.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHe added that about 20 of those who planned the overnight coup were killed and 30 more were wounded .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"killed\", \"life.die.deathcausedbyviolentevents\", \"This event represents death resulting from violent actions during a coup attempt, indicating loss of life due to conflict or aggression.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nClinton 's Republican opponent , Donald Trump Donald TrumpRepublicans and the lost promise of local control in education Press : Worse than Nixon What do Democrats stand for ?\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"promise\", \"contact.commitmentpromiseexpressintent.meet\", \"An event where one party expresses an intent to commit or make a promise during a meeting or interaction with another party.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nA 2011 report from the Government Accountability Office found that illegal immigrants and other non - citizens , in our prisons and jails together , had around 25,000 homicide arrests to their names , 25,000 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"homicide\", \"life.die.deathcausedbyviolentevents\", \"This event represents a death resulting from a violent act, specifically homicide, involving illegal immigrants and non-citizens in detention facilities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMichael Haas , administrator of the state elections division , said clerks in each of the state ’s 72 counties are responsible for programming voting equipment , and that the state ’s 1,854 municipal clerks have to conduct public tests of voting machines before polls open .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"voting\", \"inspection.sensoryobserve.monitorelection\", \"Monitoring election activities to ensure voting equipment functions correctly and securely during the electoral process.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nUnder Hillary Clinton ’s leadership as Secretary of State , the U.S. gave support to the Honduras military which ousted yet another democratically elected president .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"ousted\", \"personnel.endposition.n/a\", \"This event denotes the end of a person's official position or role in an organization or government.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe police called it sexual assault , and an appalled public called it rape .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"rape\", \"life.injure.n/a\", \"This event denotes a physical or psychological harm caused by an act of violence or aggression involving a person's bodily integrity.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWhile Clinton was right to suggest the fact - checkers get busy on his statements -- many were misleading -- if I 'm looking at who moved the needle tonight with voters , it was Trump , not Clinton .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"misleading\", \"contact.prevarication.meet\", \"An event where a person makes statements that are false or deceptive, misleading the public or audience.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe video ends with the execution of an alleged Russian intelligence agent , who is reportedly shot by one of the militants .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"execution\", \"justice.judicialconsequences.execute\", \"The event represents the judicial or extrajudicial killing of an individual, typically as a consequence of alleged criminal activity or intelligence operations.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn his research from St. Petersburg , Chen discovered that Russian internet trolls — paid by the Kremlin to spread false information on the internet — have been behind a number of \" highly coordinated campaigns \" to deceive the American public .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceive\", \"contact.prevarication.correspondence\", \"This event represents intentional misrepresentation to mislead others through false information in digital communication.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAnd you will be deceived by Clinton .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceived\", \"contact.prevarication.meet\", \"An event where one party is misled or tricked by another through false information or deception in a meeting or interaction.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThey opted to instead detonate the bomb at the base of the tomb , causing minimal damage .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"detonate\", \"conflict.attack.n/a\", \"An attack occurred without a specific target or category, indicating a general hostile action with unclear classification.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nPutin enforces draconian controls against journalists and political opposition , literally killing his enemies , like the former chief of RT who ended up bludgeoned in a Washington , DC hotel room .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"killing\", \"life.die.deathcausedbyviolentevents\", \"An individual dies as a result of violent actions, indicating intentional harm through physical force or attack.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAsked by Newsweek , “ Do you , or any of your business units have outstanding loans with Russian banks or individuals ? ” his spox said “ Mr.\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"loans\", \"transaction.transfermoney.borrowlend\", \"Represents a financial borrowing or lending action between entities involving money transfer to or from a bank or individual.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nSome onlookers have suggested that the alleged thieves could be linked to Russian intelligence agencies , citing a recent flareup of activity by \" Guccifer 2.0 , \" believed to be a Russia - sponsored puppet account , and the country 's well - known penchant for \" false flag \" operations , intended to deceive or confuse people .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceive\", \"contact.prevarication.correspondence\", \"This event captures attempts to mislead through false information or misleading communication in political or intelligence contexts.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n‘ There were no air strikes in Aleppo on the Saturday but by Sunday it was like normal again , with bombs being dropped every day .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"bombs\", \"conflict.attack.n/a\", \"Represents an attack event in a conflict zone, indicating hostile actions such as bombings without specifying the perpetrator or victim.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWhile the United States exports little food to Russia , some European Union countries had a significant trade relationship with Russia prior to the embargo .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"embargo\", \"transaction.transfermoney.embargosanction\", \"An embargo restricts trade or financial transactions between nations, often due to political or economic sanctions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTAIPEI , Taiwan — With the lifting of the US arms embargo to Vietnam , a US defense industry source indicates Hanoi is seeking to improve its air defense and maritime security capabilities with the procurement of F-16 fighter aircraft from the US Pentagon ’s excess defense articles ( EDA ) program and refurbished P-3C Orion maritime patrol aircraft , armed with torpedoes .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"embargo\", \"transaction.transfermoney.embargosanction\", \"This event represents a financial restriction imposed by one country on another, blocking trade or transactions involving specific goods or services.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nANDOVER ( CBS ) – A woman was arrested after she allegedly vandalized a resident ’s Donald Trump campaign sign and nearly ran over the homeowner as she tried to speed away .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"vandalized\", \"artifactexistence.damagedestroy.n/a\", \"An artifact is intentionally damaged or destroyed through vandalism, often involving defacement or destruction of property.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAccording to her lawyers , she was captured by the rebels on the morning of 17 June whilst trying to reach several Aidar fighters who had been wounded in an ambush .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"wounded\", \"life.injure.n/a\", \"An individual sustained physical harm during a conflict, indicating injury as a result of combat or violence in a war-related scenario.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n\" If the PYD does not retreat to east of the Euphrates , we have the right to do everything about it , \" the minister said .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"retreat\", \"conflict.yield.retreat\", \"A retreat in conflict where one side withdraws from a contested area, signaling a strategic shift in military or political positioning.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAlthough Democrats are complaining that the Project Veritas videos were edited ( as are virtually all TV videos ) , and therefore misleading , Creamer , as well as Foval , who admitted on tape to hiring “ mentally ill ” people to disrupt Trump events and foment violence , were ousted almost immediately , with Creamer stepping down and Foval being “ laid off . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"ousted\", \"personnel.endposition.firinglayoff\", \"An event indicating the removal of a person from a position through firing or layoff, reflecting organizational action against an individual.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAccording to Haaretz , the documents also show that Bank Hapoalim used the Panamanian law firm to manage activity for various trust funds until 2011 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"funds\", \"transaction.transfermoney.giftgrantprovideaid\", \"This event represents the transfer of money as a gift, grant, or form of aid between parties, indicating financial support or distribution.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nDuring one trip to Moscow , Trump bragged that they all showed up to meet him to discuss projects around the globe .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"meet\", \"contact.discussion.meet\", \"A meeting occurred where parties discussed global projects, indicating interaction and collaboration on international initiatives.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHere -- quite apart from any legitimate concern stemming from her recent illness -- are a few of the more ludicrous and resilient theories :\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"illness\", \"life.injure.n/a\", \"An event indicating a personal health impairment or physical harm, without a specified cause or agent.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nRoldugin listed German Chancellor Angela Merkel , French President Francois Hollande , and U.S. President Barack Obama as being guilty of not taking responsibility and of concealing the true nature of agreements they had struck .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"guilty\", \"justice.judicialconsequences.convict\", \"This event signifies a formal finding of guilt in a legal or moral judgment against individuals for failing to take responsibility in political agreements.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAccording to tax records , Sea Change Foundation at the time was receiving a large infusion of funds from a mysterious Bermuda - based entity called ‘ Klein , Ltd.’ … Who owns Klein ?\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"funds\", \"transaction.transaction.giftgrantprovideaid\", \"This event represents the transfer of funds from one entity to another for support or aid purposes.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nKORNACKI : In terms of the political battle that we 're seeing play out right now , the White House is saying the Republicans have no business obstructing a nominee the president puts forward .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"obstructing\", \"movement.transportperson.prevententry\", \"Preventing a person from entering a location through obstruction during transport or movement processes.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWe will enforce all of our immigration laws .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"immigration\", \"inspection.sensoryobserve.physicalinvestigateinspect\", \"An official physical check or examination of individuals or documents to verify compliance with immigration regulations.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAnnouncing the execution , Kim Jong - un denounced him as “ human scum worse than a dog ” for an alleged attempt to overthrow him and introduce reform .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"execution\", \"justice.judicialconsequences.n/a\", \"This event denotes a judicial punishment involving death, indicating state-sanctioned execution of an individual for alleged wrongdoing.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n( CNN ) When an \" overheated and dehydrated \" Hillary Clinton stumbled into a van as she left a 9/11 memorial ceremony on Sunday morning , conspiracy theorists rejoiced .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"dehydrated\", \"life.injure.illnessdegradationhungerthirst\", \"This event represents physical deterioration due to lack of water or nutrients, leading to illness or weakness in a living being.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThey 'll mutilate me and my uniform .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"mutilate\", \"life.injure.injurycausedbyviolentevents\", \"This event represents a violent act causing physical harm, specifically mutilation, to a person or object in a harmful context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHe met with Lao President Bounnhang Vorachit , was feted at a welcome banqueted , where he toasted to a relationship he said would \" mean greater progress and opportunity for the people of Laos . \"\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"met\", \"contact.discussion.meet\", \"Two individuals engaged in a formal meeting to discuss relations, policies, or mutual interests in a diplomatic or personal context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWhat kind of press environment is this going to lead to post - election ?\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"election\", \"inspection.sensoryobserve.monitorelection\", \"Monitoring an election involves observing and assessing the electoral process for fairness, transparency, and accuracy in real time.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nImagine , too , that it was the Democrat ’s presumptive White House nominee who considered appointing as her running mate a retired three - star general and former head of the Defense Intelligence Agency who last December bizarrely — and certainly ill - advisedly — chose to be a feted guest at a 10th anniversary gala in Moscow for Russia Today , the virulently anti - American Kremlin propaganda outlet and cheerleader for the Crimea land - grab of Crimea and Putin ’s bombing of U.S.-supported Syrian rebels .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"appointing\", \"personnel.startposition.hiring\", \"An organization formally assigns a person to a new official position within its leadership structure.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHowever , in a devastating critique of Australia ’s methods , medical ethicists have warned there is “ increasing evidence that Australia is engaged in torturing asylum seekers ” with refugees imprisoned for more than a year without trial .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"torturing\", \"life.injure.injurycausedbyviolentevents\", \"This event represents physical or psychological harm inflicted through violent or abusive actions on individuals, particularly in institutional settings.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn her “ Fox News Sunday ” interview , Clinton was asked about Smith ’s allegations that she misled her and another Benghazi family after the 2012 attack .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"misled\", \"contact.prevarication.correspondence\", \"An event where one party intentionally provides false information to another during communication or interaction.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWhat Clinton said was false and misleading .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"misleading\", \"contact.prevarication.meet\", \"An interaction where one party provides false or deceptive information to another, misleading their understanding.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ It is why we call for the support and strengthening of the BDS movement and for the boycott of Israeli goods . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"boycott\", \"transaction.transferownership.embargosanction\", \"This event represents the imposition of sanctions or restrictions on Israeli goods through a boycott initiative.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nSlovenia and the other NATO members share intelligence , work to stop human trafficking , monitor the flow of refugees , and help build the region ’s security forces .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"trafficking\", \"movement.transportperson.smuggleextract\", \"This event involves the illegal transport of people for exploitation, often involving coercion or deception in the context of human smuggling.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThey pointed to allegations by guards at a detention centre on the island of Nauru “ of waterboarding , familiar to most as a torture technique that simulates drowning used by the Central Intelligence Agency ( CIA ) in places like Guantanamo bay ” .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"drowning\", \"life.die.deathcausedbyviolentevents\", \"Death resulting from violent physical events such as drowning, waterboarding, or other forms of torture or assault.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBut the real truth is that , as major banks in America stopped lending him money following his many bankruptcies , the Trump organization was forced to seek financing from non - traditional institutions .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"lending\", \"transaction.transferownership.n/a\", \"This event represents a transfer of financial control or ownership where lending is discontinued or redirected to alternative sources.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nRichard Holbrooke ( who died in 2010 ) was a high - profile US diplomat , managing director of Lehman brothers , a board member of NED , CFR , the Trilateral Commission , the Bilderberg steering group and an advisor to Hillary Clinton and John Kerry .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"died\", \"life.die.nonviolentdeath\", \"This event denotes the nonviolent end of a person's life, indicating natural or peaceful death without conflict or violence.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nA source on Tuesday told Middle East Eye that the US had confirmed it would allow the two Gulf states to begin shipments .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"shipments\", \"movement.transportartifact.disperseseparate\", \"Refers to the separation and distribution of goods via transport, indicating a logistical dispersal process between entities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ As a matter of policy , we do not provide information regarding investigations . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"investigations\", \"justice.investigate.investigatecrime\", \"This event represents an official investigation into a criminal offense conducted by justice authorities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAnd just a day before his tweet about Clinton , Trump tweeted that he was \" shocked by the heinous murder \" of a 13-year - old Israeli - American Jew who was stabbed to death by a Palestinian terrorist Thursday in her home in a West Bank settlement .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"stabbed\", \"conflict.attack.stabbing\", \"A stabbing attack involving violence against an individual, typically in a conflict context, resulting in injury or death.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThen - Prime Minister Nouri al - Maliki stripped his political opponents of power , appointed his cronies to run the army , and killed peaceful protesters .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"appointed\", \"personnel.startposition.hiring\", \"This event represents the appointment of individuals to new positions within an organization or institution.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBut in practice , the government has taken back a number of areas with starve - or - surrender tactics , bombarding and starving people until they agree to leave .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"starve\", \"life.injure.illnessdegradationhungerthirst\", \"This event represents physical harm caused by hunger and thirst, leading to illness or degradation of health in individuals.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nFacebook Twitter Pinterest Cornel West urges Obama to pardon Edward Snowden .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"urges\", \"contact.commandorder.n/a\", \"An event where one individual pressures another to take a specific action, such as issuing a pardon.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThis comes just as the French prime minister has announced that his government plans to intensify its restrictions on free speech targeting the boycott , divestment and sanctions ( BDS ) movement .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"boycott\", \"transaction.transferownership.embargosanction\", \"Government imposes restrictions on free speech targeting BDS movement through embargos or sanctions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nEventually I expected him to announce he knew best how to defeat ISIS because he built a hotel in Mosul .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"built\", \"manufacture.artifact.n/a\", \"This event represents the creation of a physical object, such as a hotel, through construction or development activities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThey point to not only Sloyan ’s alleged decapitation but also the killing of three elderly residents of Talish , a village in northern Karabakh close to the Armenian - Azerbaijani “ line of contact . ” According to the Karabakh Armenian authorities , the villagers were shot dead in their home by Azerbaijani special forces that briefly infiltrated Talish early on April 2 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"shot\", \"life.die.deathcausedbyviolentevents\", \"This event captures deaths resulting from violent attacks, specifically highlighting fatalities due to gunshot wounds in conflict-related incidents.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nPreliminary findings suggest all three attackers at Ataturk airport opened fire and then detonated explosives strapped to their bodies , similar to the mass shootings and suicide bombings at Paris ' Bataclan concert hall in November .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"detonated\", \"conflict.attack.bombing\", \"An attack involving explosives detonated by attackers, combining gunfire and suicide bombing elements in a coordinated assault.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe resolution further seeks to ensure the release of those imprisoned for “ legitimate ” freedom of expression online .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"imprisoned\", \"justice.arrestjaildetain.arrestjaildetain\", \"This event represents the act of arresting or detaining individuals, often due to legal or political reasons related to expression or dissent.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIt would be better still , of course , were Sanders to address AIPAC and use the occasion , speaking as a Jew , to advocate for justice for Palestinians ; and , speaking as a candidate , to announce that , if elected , the United States will stop enabling Israel ’s efforts to ethnically cleanse Palestine .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"advocate\", \"contact.requestadvise.meet\", \"An event where one party requests advice or guidance from another through a meeting or interaction.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nFans skirmish ahead of the match in the French port city of Marseille on Saturday .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"skirmish\", \"conflict.attack.n/a\", \"A minor conflict or confrontation between individuals, often involving physical or verbal aggression in a public setting.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nShe added : “ The FBI believes the true motives of the Russian partners , who are often funded by the government , is to gain access to classified , sensitive , and emerging technology from the companies . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"funded\", \"transaction.transfermoney.giftgrantprovideaid\", \"Represents financial support given to entities, often for strategic or political purposes, involving government-backed funding.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBut Mr Wang urged the US to rethink its strategy .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"urged\", \"contact.requestadvise.meet\", \"An event where one party requests another to reconsider a strategy or course of action through direct engagement.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMoreover , finding millions of undocumented migrants almost certainly would entail immigration agents entering homes , raiding businesses and operating roadblocks to check identity papers to separate U.S. citizens and approved immigrants from those in the country illegally , a winnowing - out process that undoubtedly would be challenged in court .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"roadblocks\", \"movement.transportperson.prevententry\", \"Immigration agents block roads to stop unauthorized entry and verify identities of migrants entering the U.S.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nYou know , I have been campaigning for Democrats , fundraising for Democrats , recruiting for Democrats to run and win for a really long time , I think about 40 years , \" Clinton said .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"recruiting\", \"personnel.startposition.hiring\", \"This event signifies the initiation of hiring new personnel to hold positions within a political organization or campaign.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn recent days , Russian and Syrian aircraft have carried out attacks , mainly in Aleppo , using bunker - busting bombs , incendiary munitions , cluster bombs , barrel bombs and thermobaric bombs , which scatter a cloud of explosive particles before detonating in devastating blasts , according to American intelligence officials .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"detonating\", \"conflict.attack.n/a\", \"Represents an attack carried out during a conflict, involving military actions against a target using explosive weaponry.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nRachel Maddow points out that while Donald Trump is confidently declaring that Russia is fighting ISIS in Syria , Russia has constructed an air defense system in Syria against planes that ISIS does n't have but the U.S. does .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"constructed\", \"manufacture.artifact.n/a\", \"The creation of a physical object or system, such as a defense installation, through construction or building activities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHe found that they would tear up the living area and create chaos .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"tear\", \"artifactexistence.damagedestroy.n/a\", \"This event indicates damage or destruction of an object or area, implying disruption or ruin of a physical space or item.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWhen pressed on whether assault rifles ought to be restricted in the wake of the Orlando massacre , Gov.\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"massacre\", \"conflict.attack.n/a\", \"Represents a violent attack event with no specific classification, indicating a large-scale violent incident involving harm or death.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThis is why they voted for Obama in 2008 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"voted\", \"government.vote.n/a\", \"This event represents a vote cast by individuals in a government election, specifically for a presidential candidate in 2008.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe one common mechanical process in almost every aspect of American politics is the two - party system : an extra - constitutional artifice that long ago hijacked government .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"hijacked\", \"conflict.attack.stealrobhijack\", \"This event represents the unauthorized takeover of control or assets through force or deception in political systems.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nComey at the time called Clinton \" extremely careless \" but declined to recommend criminal charges to the Justice Department , outraging Republicans who said she should have been indicted .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"indicted\", \"justice.initiatejudicialprocess.n/a\", \"Initiating a judicial process to formally charge someone with a crime through indictment by a grand jury or prosecutor.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHillary Clinton , Stein agrees with me , has considerably less capacity to deceive and bamboozle progressive and young voters than Barack Obama enjoyed in 2007 - 08 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceive\", \"contact.prevarication.n/a\", \"This event captures an instance where someone attempts to mislead or manipulate others through false statements or deception.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn the memo , the State Department officials wrote that the Assad government ’s continuing violations of the partial cease - fire , known as a cessation of hostilities , will doom efforts to broker a political settlement because Mr. Assad will feel no pressure to negotiate with the moderate opposition or other factions fighting him .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"negotiate\", \"contact.discussion.correspondence\", \"This event represents official communication about discussions aimed at reaching an agreement between parties in a political conflict.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHide Caption 30 of 83 Photos : Inside North Korea Vendors sell flowers February 15 to mourners paying their respects to deceased leaders of North Korea .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceased\", \"life.die.nonviolentdeath\", \"An individual dies without violence, typically due to natural causes or illness, indicating a peaceful end to life.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nOther key points of the resolution include investigating attacks against bloggers or other internet users , and refraining from preventing access to information online by , for example , shutting down the internet during key times such as elections or terror attacks .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"investigating\", \"justice.investigate.investigatecrime\", \"This event involves probing illegal activities targeting bloggers or internet users through formal investigations by authorities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHe told local television he was seeking asylum because of persecution by the Russian security service .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"asylum\", \"movement.transportperson.grantentryasylum\", \"An individual is granted asylum entry due to persecution by a state's security forces, indicating legal protection from harm.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAn inspection team led by Independent Chair Rune Andersen , an independent international anti - doping expert ( Norwegian ) and three members of the IAAF Council who will be appointed in the next few days . ’\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"inspection\", \"inspection.sensoryobserve.physicalinvestigateinspect\", \"An official physical examination of facilities or practices to verify compliance with anti-doping regulations and standards.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBut Booth said an arms embargo could only work if neighboring countries cooperated .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"embargo\", \"transaction.transfermoney.embargosanction\", \"An embargo restriction on arms trade enforced through international cooperation to limit weapon flow between nations.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAccording to the sources , who spoke on condition of anonymity , the slain military officials included four Russian generals and four Syrian generals who had met late Tuesday in Turkmen Mountain -- located in Syria ’s northern Latakia region -- to discuss recent military developments .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"slain\", \"life.die.deathcausedbyviolentevents\", \"This event indicates that individuals died due to violent actions, specifically in a military context involving armed officials.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nREAD MORE : Pope urges reconciliation during trip to Sarajevo\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"urges\", \"contact.commandorder.broadcast\", \"Pope issues a call for reconciliation during a visit to Sarajevo, promoting peace and unity among divided communities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThey ’re devastated by the smallness of the catch , and they ’re butting up against the six million victims of the Holocaust .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"Holocaust\", \"life.die.deathcausedbyviolentevents\", \"Death resulting from violent events such as war, genocide, or mass violence where individuals lose their lives due to systemic or large-scale conflict.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMuch of what is known emerged only after Yanukovych fled Ukraine in 2014 , ousted by a popular revolution , and volunteer divers retrieved thousands of internal government documents that had been hastily thrown into a lake at his presidential retreat .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"ousted\", \"personnel.endposition.firinglayoff\", \"This event denotes the removal of a person from a position due to political or social pressure, often through public revolt or protest.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIt appears , for example , that none of the anti - Muslim critics are aware that , with respect to the single greatest source of deadly intentional violence worldwide , Muslim societies are among the least violent in the world .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"violent\", \"life.injure.n/a\", \"This event denotes a situation where life is injured without a specific cause or agent identified in the context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n1/6 A woman walks through rubble from a building destroyed by shelling from forces loyal to Syrian President Bashar Al - Assad in downtown Aleppo REUTERS / Goran Tomasevic\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"walks\", \"movement.transportperson.selfmotion\", \"Describes a person moving through an environment, indicating physical movement without external transport or vehicle use.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nSaudi Arabia has perpetrated a mass execution that puts ISIS ’s beach beheadings to shame .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"execution\", \"justice.judicialconsequences.n/a\", \"This event type denotes a state-sanctioned killing, often involving legal or political implications, indicating severe judicial or governmental action.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMrs. Clinton is by far the more conventionally experienced and credentialed candidate , and in some respects is more hawkish : She has called for the creation of a no - fly zone in Syria , as well as more airstrikes there , and has urged closer collaboration with Silicon Valley to expand the United States ’ surveillance capabilities , alarming some civil liberties advocates .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"urged\", \"contact.requestadvise.broadcast\", \"The event represents a request for advice or collaboration through public communication, indicating a formal outreach to influence policy or strategy.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ Ever since the June 28 , 2009 , coup that deposed Honduras ’s democratically elected president , José Manuel Zelaya , the country has been descending deeper into a human rights and security abyss .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"coup\", \"conflict.attack.n/a\", \"A sudden violent takeover of government power, often involving force, leading to instability and human rights violations.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nOrganizations have repeatedly called for boycotts in the world , of Myanmar ( Burma ) , Russia , China or Mexico without ever invoking this clause .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"boycotts\", \"transaction.transferownership.embargosanction\", \"This event represents a formal restriction on trade or activities imposed on a country or entity through international boycotts or sanctions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAlthough Snowden ’s actions aided my push for reform , the fact remains that Snowden broke an oath he willingly took to protect our national security and classified secrets .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"oath\", \"contact.commitmentpromiseexpressintent.meet\", \"An event where an individual expresses a formal commitment or promise to uphold a duty or obligation, such as a sworn oath to protect national security.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nREAD MORE : Armenian genocide : 130 K march in LA to mark 100th anniversary ( PHOTO , VIDEO )\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"genocide\", \"life.die.deathcausedbyviolentevents\", \"This event represents death resulting from violent conflicts or mass atrocities, such as genocide or war-related violence.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n+ show all 10 things immigration has done for Britain\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"immigration\", \"inspection.sensoryobserve.inspectpeopleorganization\", \"Involves observing or inspecting people or organizations through sensory means to assess their characteristics or compliance.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nImage copyright AP Image caption IS proclaimed the creation of a caliphate stretching across Syria and Iraq in 2014\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"proclaimed\", \"contact.publicstatementinperson.n/a\", \"An individual or group makes a public declaration in person, asserting a fact or claim about a situation or entity.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAnd they may also want to look into how Abedin failed to turn over husband Anthony Weiner ’s laptop containing thousands of Clinton emails — after she swore under oath that she had handed over “ all the devices ” she possessed containing such emails .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"oath\", \"contact.commitmentpromiseexpressintent.meet\", \"An event where one party expresses intent to meet or commit to a promise through a formal commitment under oath.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nSweden , which professed to be neutral in World War II , was a critical supplier of high - quality munitions to the Nazis , permitting them to cross Swedish territory to invade Norway , and usually rejecting refugees .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"professed\", \"contact.publicstatementinperson.n/a\", \"A public declaration of neutrality made in person by a nation during wartime, indicating official stance despite actions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe worst abuses are rooted in the so - called microfinance industry , which extends small loans to tide borrowers over until their next paycheck .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"loans\", \"transaction.transferownership.borrowlend\", \"Represents a financial transaction where one party lends small amounts to another for short-term use until next income.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe Potsdam Declaration , announced 10 days later , called for Japan 's unconditional surrender , threatening \" prompt and utter destruction . \"\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"surrender\", \"conflict.yield.n/a\", \"This event signifies a conflict outcome where one side yields under pressure, often due to threats or strategic disadvantages.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nReporters who investigate these ties are being very careful about their conclusions , but this looks really bad .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"investigate\", \"justice.investigate.investigatecrime\", \"Reporters are conducting a careful investigation into potential criminal ties involving individuals or organizations.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe release of a video of that shooting set off major protests in the city , prompted the firing of the department 's top cop and the U.S. Justice Department 's launch of a civil rights investigation of Chicago police .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"firing\", \"personnel.endposition.firinglayoff\", \"The removal of a public official from their position due to misconduct or performance issues, resulting in organizational change.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nFalsely accused Clinton of laughing at a 12-year - old rape victim : “ Her client she represented got him off and she 's seen laughing at the girl who was raped . ” ( Clinton laughed at various points of a 1980s interview about the case – but never at the victim .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"raped\", \"life.injure.n/a\", \"An injury event involving harm to a person's physical or mental well-being due to violent or harmful actions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThey need large forest tracts to roam in . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"roam\", \"movement.transportperson.selfmotion\", \"An individual moves independently through a natural environment, such as a forest, for exploration or survival purposes.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAssange remains in exile in the Ecuadorian Embassy in London to avoid prosecution on sexual assault charges in Sweden .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"assault\", \"life.injure.n/a\", \"This event denotes a physical or psychological harm inflicted on an individual through an assault incident, involving illegal or violent contact.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn his research from St. Petersburg , Chen discovered that Russian internet trolls — paid by the Kremlin to spread false information on the internet — have been behind a number of \" highly coordinated campaigns \" to deceive the American public .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"deceive\", \"contact.prevarication.n/a\", \"This event represents intentional misrepresentation to mislead others, involving deception through false information dissemination.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe claims by Hezbollah sources come even as the Gulf Cooperation Council ( GCC ) on Wednesday declared the group a terrorist organization , citing Hezbollah 's activity in the Gulf states recruiting terrorists and smuggling weapons - a point that was not raised in the article .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"recruiting\", \"personnel.startposition.hiring\", \"This event represents the initiation of a personnel role through hiring, indicating a new employment relationship established within an organization.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nEarlier this year , Stein made controversial comments suggesting Wi - Fi is harming students .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"harming\", \"life.injure.n/a\", \"This event indicates a harmful impact on a living being, potentially leading to physical or psychological damage due to exposure to a harmful factor.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThey range from the slaughter of hundreds of thousands of Filipinos resisting U.S. imperial invasion and occupation at the last century ’s outset to the restoration of de facto slavery in Haiti and the Dominican Republic after World War I ; the unnecessary atom bombing of Hiroshima and , even worse , of Nagasaki ( really the first shots of the Cold War ) ; the toppling of more than 50 governments by U.S. coups and invasions since the end of World War II ; the liquidation of perhaps as many as 5 million Southeast Asians in the so - called Vietnam War between 1962 and 1975 ; the Cold War - era sponsorship of Third World fascism from Chile to South Africa and Indonesia ; the attempted assassinations of Fidel Castro and numerous CIA - directed terror bombings in socialist Cuba ; the near instigation of global thermonuclear war on at least three occasions ; the development and sponsorship of Osama bin Laden and other radically arch - reactionary , jihadist Muslim , paramilitary forces to fight the Cold War against the Soviet Union in Afghanistan ; the “ Highway of Death , ” when U.S. warplanes engaged in an aerial traffic jam as they rushed to slaughter tens of thousands of surrendered Iraqi troops retreating from Iraq in 1991 ; the coordination and sponsorship of a mass - murderous civil war on peasants , workers and intellectuals ( with a death toll well into the many hundreds of thousands ) in Central America during the 1970s and 1980s ; the disastrous U.S. invasion of Iraq ( responsible for at least 1 million Iraqi deaths ) ; the calamitous U.S. toppling of the Libyan Gadhafi regime ; the calamitous destabilization of the Syrian regime ; the U.S. funding and encouragement of civil war in central Africa ; the enablement and protection of a vicious right - wing coup in Honduras in the spring and summer of 2009 ; the criminal U.S. global war of terror , replete with rampant “ targeted assassinations , ” torture , illegal renditions , endless drone war and special - forces killing operations across the Muslim world and other places as well .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"retreating\", \"conflict.yield.n/a\", \"This event type signifies a conflict outcome where forces engage in combat against retreating opponents, resulting in significant casualties during military retreats.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ I was the only person to hold trials , and not just shoot people .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"shoot\", \"life.injure.injurycausedbyviolentevents\", \"This event captures injuries resulting from violent actions, such as shootings or attacks, involving physical harm to individuals.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nDespite Tuesday 's development , Ecuador will continue to provide asylum to Assange , according to the statement .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"asylum\", \"movement.transportperson.grantentryasylum\", \"An individual is granted asylum entry, allowing safe passage and protection in a foreign country due to political or safety concerns.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ We are disappointed at the DCCC ’s placating response to our demand to value all Black life , ” said the statement .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"demand\", \"contact.commandorder.broadcast\", \"An event where a group issues a formal request or instruction to another entity, signaling a directive or expectation.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nYou 'd have to either register using a false name in multiple places and do this on a large enough scale to affect the outcome or you 'd have to hire people to go into the polling place , claim that they were someone else on the voting rolls , hope those people on the voting rolls had n't voted yet and then vote for them and that this scheme could be done on a large enough scale to affect the outcome of an election .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"voting\", \"government.vote.n/a\", \"This event represents an unauthorized voting action that could influence election outcomes through fraudulent or illegal means.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHistorically , as Republicans positioned themselves alongside soldiers and cops , Democrats favored military cutbacks and were more prone to protest in the streets , creating an image as just a little less red , white and blue .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"protest\", \"conflict.demonstrate.marchprotestpoliticalgathering\", \"A political protest or demonstration involving public gathering to express opposition or support for policies, often with a political or ideological motive.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWhen pressed on whether assault rifles ought to be restricted in the wake of the Orlando massacre , Gov.\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"massacre\", \"life.die.deathcausedbyviolentevents\", \"This event represents a large-scale violent death caused by an attack, involving multiple fatalities in a single incident.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAlthough Democrats are complaining that the Project Veritas videos were edited ( as are virtually all TV videos ) , and therefore misleading , Creamer , as well as Foval , who admitted on tape to hiring “ mentally ill ” people to disrupt Trump events and foment violence , were ousted almost immediately , with Creamer stepping down and Foval being “ laid off . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"ousted\", \"personnel.endposition.n/a\", \"This event indicates the termination of a person's position in an organization, often due to misconduct or controversy.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nUnsurprisingly , Clinton acquiesced and chose one of the officials recommended by Soros — Miroslav Lajcak .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"recommended\", \"contact.requestadvise.correspondence\", \"An official requests advice through correspondence, indicating a formal communication for guidance or recommendation in decision-making.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe deal , which required that Assad halt airstrikes against civilian areas in places like Aleppo , quickly collapsed and the U.S. and Russia have cut off formal dialogue about Syria .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"dialogue\", \"contact.discussion.correspondence\", \"A formal exchange of views between parties on a specific issue, often involving negotiation or discussion of policies or agreements.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nSo he proved the rule by making a calculated and exceptional decision : he opted for a Trojan Horse strategy , hoping to invade and conquer the duopoly from within .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"invade\", \"conflict.attack.n/a\", \"An attack event where one party initiates hostile action against another, often involving force or strategic deception.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThey were civilian massacres . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"massacres\", \"conflict.attack.n/a\", \"Represents a violent attack on civilians, involving mass killing without clear combat context or defined combatants.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nUS secretary of state John Kerry said he had spoken to China ’s foreign minister Wang Yi and hoped that Beijing would continue to cooperate with the United States UN sanctions aimed at rolling back North Korea ’s nuclear weapons program .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"sanctions\", \"transaction.transaction.embargosanction\", \"Sanctions imposed by a state to restrict activities of another, often to pressure policy changes or halt illicit actions like nuclear development.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThey were civilian massacres . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"massacres\", \"life.die.deathcausedbyviolentevents\", \"This event represents civilian deaths resulting from violent attacks or massacres involving large-scale killings of non-combatants.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nKnowing Hillary Clinton ’s position of being for the death penalty , she wanted to give a heads up , writing “ Here ’s one that worries me about HRC ” ( Hillary Rodham Clinton ) .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"death penalty\", \"justice.judicialconsequences.execute\", \"This event signifies the potential execution of a person via state-sanctioned lethal punishment as a judicial consequence.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAround this time , the plane made a 90-degree turn to the left and then a full circle to the right , dropping precipitously to 15,000 feet from 37,000 and then plunging again to 9,000 feet before it disappeared from radar .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"plunging\", \"movement.transportartifact.fall\", \"The event represents a rapid descent of a transport artifact, indicating a sudden and dangerous drop in altitude during flight.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAfter the embargo began in 2014 , Valio sold its \" Oltermanni \" cheese , popular in Russia , for half its regular price in Finland because of oversupply .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"embargo\", \"transaction.transferownership.embargosanction\", \"An ownership transfer occurs due to an embargo, where assets are sold or restricted under international sanctions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe primary goal , however , is to provide aid and protection to refugees in the countries and areas where they have fled , “ so refugees can return home when the conflict ends , ” the official added .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"provide aid\", \"transaction.transaction.giftgrantprovideaid\", \"This event represents the act of giving aid or resources to refugees in displacement areas to support their return after conflict ends.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIt had been designed to carry entire military units and all their equipment during the Vietnam War and Desert Storm .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"War\", \"conflict.attack.n/a\", \"Represents a conflict-related attack with no specific classification, indicating an unspecified military engagement during wartime operations.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n\" The occasion of this campaign should be seized not only to expose the lies of Hillary Clinton , but also to seek freedom from America 's seven decades of subjugation to the military - industrial complex and its organic intellectuals who never cease conjuring up ' threats ' and ' enemies ' to justify the war economy , \" the journalist observed .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"lies\", \"contact.prevarication.broadcast\", \"This event captures the act of spreading false claims about a political figure through public communication.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n… [ Russian media ] depict the West as grim , divided , brutal , decadent , overrun with violent immigrants and unstable .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"violent\", \"life.injure.injurycausedbyviolentevents\", \"An injury resulting from violent events involving physical or psychological harm caused by aggression or conflict.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMrs Clinton said the other emails are private - relating to topics like her daughter 's wedding , her mother 's funeral and \" yoga routines \" .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"funeral\", \"contact.funeralvigil.n/a\", \"An event marking a funeral vigil, involving mourning and remembrance of a deceased person in a formal gathering.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThey claim that all of the named plaintiffs already knew that the DNC was biased when they donated — so therefore how could they have been duped if they knew ?\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"donated\", \"transaction.transfermoney.giftgrantprovideaid\", \"An event where money or aid is transferred as a gift, grant, or support to another party, often involving financial or institutional contributions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nA region where small skirmishes can trigger big wars would turn into a nuclear tinderbox . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"skirmishes\", \"conflict.attack.n/a\", \"A small-scale conflict can escalate into a large-scale war due to instability and potential for escalation.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTurkey has spent much of this year reeling from terror attacks as it weathers bombing campaigns by both ISIS and Kurdish militants .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"attacks\", \"life.injure.injurycausedbyviolentevents\", \"Violent attacks result in physical harm to individuals, often due to terrorism or armed conflict in a specific location and time.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBut the French law , which includes criminal penalties , is arguably harsher than Israel ’s , which allows boycott supporters to be pursued for financial damages , but not jailed .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"jailed\", \"justice.arrestjaildetain.arrestjaildetain\", \"This event denotes the legal act of arresting and detaining someone for a crime, involving imprisonment as a penalty.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nEstimates of civilian deaths from military action , crimes against humanity , starvation and disease are as high as 9 million .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"starvation\", \"life.injure.n/a\", \"This event represents civilian harm due to indirect causes like starvation, disease, or military actions, indicating widespread suffering.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe crown prince of Bahrain , for instance , was a donor — but should n’t the secretary of state be meeting with foreign leaders anyway ?\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"meeting\", \"contact.discussion.correspondence\", \"An event where two or more parties engage in a discussion or conversation, either in person or through correspondence.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nShe borrowed $ 75 from the small blue and green kiosk of a microfinance company called Instant Money near a bus stop in Iskitim , figuring she could repay it within a month .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"borrowed\", \"transaction.transfermoney.borrowlend\", \"An individual receives funds from a lender with an agreement to repay within a specified time frame.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nShe mused , “ Having to cope , run the trailer court , even threaten to shoot a dog”—her tenant ’s pet had endangered children—”it ’s hardened me , made me act like a man .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"shoot\", \"conflict.attack.firearmattack\", \"An attack involving a firearm, indicating a violent confrontation where a weapon is used to harm or threaten someone.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nUnlike Trump , Sanders never contended , as his liberal critics suggested , that if he were elected president , he could enact the changes he advocated .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"suggested\", \"contact.requestadvise.broadcast\", \"An event where one party proposes or implies a viewpoint or claim to another party through communication.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAnd they ripped Trump for his newest controversy : urging earlier Wednesday that Russia recover and release Hillary Clinton ’s private emails .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"urging\", \"contact.requestadvise.n/a\", \"An entity requests advice or action from another, indicating a call for guidance or intervention in a specific situation.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHe urged the federal government to “ pause , ” and put in place stringent background checks before allowing Syrians to enter the U.S.\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"urged\", \"contact.requestadvise.meet\", \"An official urged a government to pause and implement policies before allowing entry of a specific group into the country.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe growing rift with the West , as well as fears over Syria and economic concerns , spurred Mr Erdogan to seek to restore relations with Moscow that were wrecked last November when Turkey shot down a Russian fighter jet over the Syrian border .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"shot\", \"life.die.deathcausedbyviolentevents\", \"An individual dies due to violent action involving a weapon, indicating a fatal outcome from physical aggression or combat.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAccording to data provided by the Senate Subcommittee on Immigration , and the national interest between 9/11 and the end of 2014 , at least 380 foreign born individuals were convicted in terror cases inside the United States .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"convicted\", \"justice.judicialconsequences.convict\", \"An individual is formally found guilty of a crime by a judicial body, resulting in legal penalties and consequences.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn it , voters who backed the Green Party candidate in 2000 voiced regret that they might have contributed to Republican George W. Bush ’s victory over Democrat Al Gore Al GorePence calls hiring of counsel ' routine ' CORRECTED : Three members of Mueller 's team have donated to Democrats OPINION : Do n't repeat the mistakes of Clinton and 1998 with Trump MORE .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"donated\", \"transaction.transferownership.n/a\", \"This event represents a transfer of ownership interest, typically involving financial or political contributions from one party to another.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWe should resurrect the old definitions of democracy that emerged from the movements of workers in struggle -- and that means organizing at the grassroots to demand social change and building for a political alternative independent of the two - party status quo .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"demand\", \"contact.commandorder.n/a\", \"This event represents a directive to require or insist upon a specific action or outcome from a group or authority.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHe has called for using torture against terrorism suspects , and has said America has no standing to lecture other nations on human rights and the rule of law , as administrations have done since the depths of the Cold War .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"torture\", \"life.injure.injurycausedbyviolentevents\", \"This event represents physical or psychological harm caused by violent actions, such as torture, inflicted on individuals during conflict or detention.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ But then again , they may also be all too aware that in exercising their duty of care they may simply be ‘ patching up ’ the victims in order for them to be tortured again . ”\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"tortured\", \"life.injure.n/a\", \"This event represents physical or psychological harm inflicted on individuals, indicating potential abuse or negligence in care duties.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nSpotted on Capitol Hill : Helen Mirren ( watch her testimony on Holocaust art restitution here ) :\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"Holocaust\", \"life.die.deathcausedbyviolentevents\", \"This event represents death resulting from violent acts, indicating a person's demise due to physical aggression or conflict.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThough the Scandinavians posed as the moral arbiters of the West , their “ model democracy ” has created many unacknowledged social problems .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"created\", \"manufacture.artifact.n/a\", \"This event type indicates the creation of an object or product without a defined category or classification in the system.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn a terse press conference Wednesday morning at his property in Doral , Florida , Donald Trump said he hopes that Hillary Clinton ’s “ lost ” emails have fallen into Russian hands and urged the foreign government to release them .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"urged\", \"contact.requestadvise.correspondence\", \"An event where one party requests or urges another to take action, such as sharing information or making a decision.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n( As he complains about the media response , he should remember what happened when Bill Clinton baldly lied about such a question : He was impeached . )\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"lied\", \"contact.prevarication.n/a\", \"An act of deliberate falsehood told to mislead, involving intentional deception in a public or official context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThis comes just as the French prime minister has announced that his government plans to intensify its restrictions on free speech targeting the boycott , divestment and sanctions ( BDS ) movement .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"boycott\", \"transaction.transferownership.n/a\", \"This event represents a transfer of ownership rights, though the specific nature or context is unspecified in the provided data.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHe could be placed in Egypt during the revolution , meeting with Wael Ghonim , the Google employee whose arrest and imprisonment hours later would make him a PR - friendly symbol of the uprising in the Western press .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"arrest\", \"justice.arrestjaildetain.arrestjaildetain\", \"An individual is taken into custody by authorities, often leading to imprisonment or detention during political or social unrest.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWe will terminate the Obama administration 's deadly , and it is deadly , non - enforcement policies that allow thousands of criminal aliens to freely roam our streets , walk around , do whatever they want to do , crime all over the place .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"roam\", \"movement.transportperson.selfmotion\", \"Describes an individual's autonomous movement through a physical space without external guidance or direction.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n“ Eventually , you will get sick , ” Rauch writes .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"sick\", \"life.injure.n/a\", \"This event indicates a personal health decline due to illness, reflecting a natural life process without specific causation.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nFREQUENT FLYER : Bill Clinton took far more rides on convicted sex offender Jeffrey Epstein 's ' Lolita Express ' airplane than has been previously disclosed , according to a Fox News report\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"convicted\", \"justice.judicialconsequences.convict\", \"This event indicates a person has been legally found guilty of a crime and faces judicial consequences as a result.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThis boycott movement is enjoying growing success around the world as the only nonviolent means to put pressure on Israel .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"boycott\", \"transaction.transaction.embargosanction\", \"An embargo or sanction imposed through a boycott to exert economic or political pressure on a target entity without violence.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nOne of the main reasons we have so little democracy today , even relative to the recent past , is that protest movements are weak , having been greatly hurt in turn by recent assaults on democratic rights : Occupy encampments ruthlessly evicted ( mostly by Democratic Party mayors , by the way ) , Verizon strikers barred from picketing in important locations , laws that make boycotts in solidarity with Palestine illegal .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"boycotts\", \"transaction.transfermoney.embargosanction\", \"This event represents state-imposed restrictions on economic activities through boycotts, limiting trade or solidarity actions against specific entities or causes.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nSouth Africa expressed its displeasure last year when it was accused of ignoring its treaty obligations because it allowed Sudan ’s President Omar Bashir — accused of genocide and other crimes in his country ’s Darfur region — to attend an African Union summit without arresting him .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"genocide\", \"life.die.deathcausedbyviolentevents\", \"Death resulting from violent acts or conflicts, indicating a loss of life due to aggression or warfare.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe candidate ’s real estate empire , unable to borrow from most U.S. banks , gets capital from Russian sources .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"borrow\", \"transaction.transfermoney.borrowlend\", \"An entity borrows funds from another, involving a financial transaction where money is transferred under loan terms.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMedia playback is unsupported on your device Media caption Stratfor analyst Sim Tack : \" This was indeed an IS ( Islamic State ) attack , rather than an accidental explosion \"\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"attack\", \"life.injure.injurycausedbyviolentevents\", \"An injury resulting from violent acts such as attacks or explosions, indicating harm caused by intentional or accidental force.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nRepresentatives of the collection industry either accused her of lying or suggested that a few “ black collectors ” were soiling the entire industry .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"lying\", \"contact.prevarication.meet\", \"An interaction where one party accuses another of lying during a contact, indicating a claim of deception in communication.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWhether it 's dangerous materials being smuggled across the border , terrorists entering on visas or Americans losing their jobs to foreign workers , these are the problems we must now focus on fixing .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"smuggled\", \"movement.transportperson.hide\", \"Involves individuals concealing themselves during transport across borders, often to evade detection or authorities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTorch procession : This was the police guard for the Olympic Torch Relay - which caused the traffic jam in which a motorcycle robbery went fatally wrong for the perpetrator\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"robbery\", \"conflict.attack.stealrobhijack\", \"A violent theft involving force or stealth, resulting in harm or loss of property during a crime event.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nFrom his talk of \" rigged elections \" to his suggestion that we need Russia to hack Clinton ’s emails if the Justice Department wo n’t indict her , Trump ’s message is clear : Clinton is so corrupt , and the system that favors her is so broken , that our ordinary democratic and legal processes are n’t equipped to handle it .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"indict\", \"justice.initiatejudicialprocess.chargeindict\", \"The event represents the initiation of a legal process to formally charge an individual with a crime through indictment by a judicial body.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nOne of the main reasons we have so little democracy today , even relative to the recent past , is that protest movements are weak , having been greatly hurt in turn by recent assaults on democratic rights : Occupy encampments ruthlessly evicted ( mostly by Democratic Party mayors , by the way ) , Verizon strikers barred from picketing in important locations , laws that make boycotts in solidarity with Palestine illegal .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"boycotts\", \"transaction.transaction.embargosanction\", \"This event represents legal restrictions on boycotts, limiting collective actions in solidarity with specific groups or causes.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTrump ’s talk of an Apple boycott after the company refused to help the FBI break into its phone was also deeply unpopular in the Valley .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"boycott\", \"transaction.transaction.embargosanction\", \"An official decision to restrict or prohibit trade or business activities with a specific entity or country, often due to political or economic reasons.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBut three years later a memo written by David Watkins to the White House chief of staff , recounting the history of the firings , suddenly surfaced .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"firings\", \"personnel.endposition.firinglayoff\", \"This event denotes the termination of a person's job position through dismissal or layoff within an organizational structure.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAccording to the company they had been abducted by “ Arab men ” who had her blindfolded , they kidnapped in an apartment and they were detained for 30 hours against their will and repeatedly raped .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"detained\", \"movement.transportperson.preventexit\", \"Preventing someone from leaving a location through forced detention or restriction of movement.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe monument to fallen Soviet sailors in the northern Latvian town of Limbazi , located about 100 kilometers from Riga , was demolished by activists from neo - Nazi group Daugavas Vanagi ( the Daugava Hawks ) , which calls itself an organization of the former Nazi Waffen SS Legion veterans .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"demolished\", \"artifactexistence.damagedestroy.n/a\", \"An artifact is damaged or destroyed, indicating a loss or removal of a physical object of historical or cultural significance.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHe also invoked Pope Francis , Turkish scholars and other figures who have recognized the massacre as genocide .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"massacre\", \"conflict.attack.n/a\", \"Represents a violent act involving harm or destruction, often tied to conflict or historical atrocities, with no specific attack category defined.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nSyrian government forces , backed by Russian air strikes , have almost encircled rebels in parts of the northern city of Aleppo .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"strikes\", \"conflict.attack.n/a\", \"Represents an attack during a conflict, involving force exerted against opposing forces in a military or political struggle.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nMore than 100 prominent physicians , bioethicists and scientists from around the world posted a letter Friday urging World Health Organization Director - General Margaret Chan to exert pressure on Olympic authorities to move the Olympics from Rio de Janeiro or delay the Games because of public health concerns over the Zika virus .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"urging\", \"contact.commandorder.correspondence\", \"A formal request sent through correspondence to influence a decision, typically involving pressure on an authority to take action.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe Wall Street Journal ran a full - page ad on Wednesday containing links to a Turkish project that denies the Armenian genocide – the 1915 massacre of 1.5 million people by Ottoman Turks .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"massacre\", \"life.die.deathcausedbyviolentevents\", \"This event represents a large-scale killing caused by violent actions, typically involving mass harm to individuals in a specific location and time.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTrump said the the Khans had \" viciously attacked \" him at the Democratic convention and then complained that they continued to assail his candidacy in a series of television interviews in recent days .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"assail\", \"conflict.attack.n/a\", \"An attack event where one party aggressively targets another, expressing hostility or harm in words or actions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nHopes flickered among some anti - Trump Republicans that there would be a revolt against him when delegates convene to nominate him formally in Cleveland from July 18 to 21 .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"convene\", \"contact.discussion.correspondence\", \"An event where individuals gather to discuss or exchange information formally, such as political delegates meeting to nominate a candidate.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThose loans average $ 125 , according to United Credit Bureau statistics .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"loans\", \"transaction.transferownership.borrowlend\", \"Represents a financial borrowing event where one party lends money to another in exchange for repayment.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nTrump has staked out a staunchly pro - Israel position , vowing in March that \" the days of treating Israel like a second - class citizen will end on day one \" of his presidency .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"vowing\", \"contact.commitmentpromiseexpressintent.broadcast\", \"An official declaration of intent to support a cause or policy, publicly communicated through speech or statement.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nA couple of weeks ago , I reposted a picture on Twitter of a guy from Florida who had taken his pickup truck and put some chain link fencing on the back of the pickup truck and claimed it was a cage that he was going to keep all the vote fraudsters in .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"vote\", \"government.vote.violationspreventvote\", \"This event indicates actions taken to prevent vote fraud by addressing alleged violations in the electoral process.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWhile railing against Asian , Latin American and Middle Eastern countries , Mr. Trump has continually praised Mr. Putin ’s government : He has hailed Mr. Putin ’s tight control over Russian society , hinted that he may not defend the NATO - aligned Baltic nations formerly in Moscow ’s sphere of influence , and for a time employed a campaign chief with close ties to Ukraine ’s pro - Russian forces .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"employed\", \"personnel.startposition.hiring\", \"This event signifies the appointment of an individual to a new position within an organization or government structure.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nA new poll showed Trump , whose unfiltered speaking style has repeatedly landed him in hot water , losing ground in three crucial states ahead of the Nov. 8 general election against Clinton .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"election\", \"inspection.sensoryobserve.monitorelection\", \"Monitoring an election involves observing voter trends, candidate performance, and public sentiment leading up to a general election vote.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe bald fact is that Trump , if he becomes president , does not know whom to appoint in order to have the support from his government to effect the changes for which his supporters hope he stands .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"appoint\", \"personnel.startposition.hiring\", \"This event represents the initiation of a personnel appointment to a new governmental position within a political administration.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nNor does Trump appear to be going so far as to say he wants to withdraw from the global system—“To all our friends and allies , I say America is going to be strong again .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"withdraw\", \"conflict.yield.n/a\", \"This event type indicates a potential reduction in engagement with global systems, suggesting a shift in international policy or alignment.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nFrom left , Melania Trump , Ivanka Trump , Mike Pence , Karen Pence , retired Gen. Michael Flynn , and former New York Mayor Rudy Giuliani watch the debate\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"debate\", \"contact.discussion.meet\", \"Individuals gather to engage in a structured conversation or exchange of views during a public debate event.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nOther fatalities included two Iraqis , one Tunisian , one Chinese , one Iranian , one Ukrainian , one Jordanian and one person from Uzbekistan , a Turkish official said .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"fatalities\", \"life.die.deathcausedbyviolentevents\", \"Records instances where individuals died due to violent events, indicating loss of life from conflict or aggression.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWe will issue detainers for illegal immigrants who are arrested for any crime whatsoever , and they will be placed into immediate removal proceedings if we even have to do that .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"arrested\", \"justice.arrestjaildetain.arrestjaildetain\", \"An arrest and detention of illegal immigrants for any crime, leading to immediate removal proceedings.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe CNAS paper urges Congress to pass the TPP as soon as possible , presumably before the election , to strengthen the U.S. “ stewardship ” of Asia .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"urges\", \"contact.commandorder.n/a\", \"An event where a party issues a directive or recommendation to another entity to take a specific action promptly.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nReports emerged last month that Cruz did not disclose loans from the financial firm during his 2012 Senate campaign .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"loans\", \"transaction.transferownership.borrowlend\", \"Represents a financial borrowing or lending transaction where one party lends money to another, involving ownership transfer of funds.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nBiałowieża hosts Europe ’s largest bison population and wolves and lynx still roam freely across its sun - mottled interior .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"roam\", \"movement.transportperson.selfmotion\", \"Describes self-movement of individuals across a geographical area without external transport.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nAccording to the Ministry of Foreign Affairs ( MoFA ) for the Republic of Turkey , the two captured CIA pilots were not only responsible for shooting down the Russian Sukhoi Su-24 m bomber over Syria last year , but also attempted to kill President Erdogan during the failed “ coup ” attempt a few days ago .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"shooting\", \"life.injure.injurycausedbyviolentevents\", \"This event captures injuries resulting from violent actions, such as shootings, involving human beings in conflict or hostile situations.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIn doing so , she was implicitly rebuking her rival , Republican nominee Donald Trump , who has questioned the need for the Western alliance and suggested that if he is elected president , the United States might not honour its NATO military commitments , in particular regarding former Soviet republics in the Baltics .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"suggested\", \"contact.requestadvise.broadcast\", \"An event where one party proposes a suggestion or opinion regarding a policy or action, often influencing others' perspectives or decisions.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nCoe said : ‘ Today we have been dealing with the failure of ARAF and made the decision to provisionally suspend them , the toughest sanction we can apply at this time .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"sanction\", \"transaction.transferownership.embargosanction\", \"This event represents the temporary suspension of an entity due to failure, applying the strictest available sanction against it.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\n\" Mr. Murphy has also challenged Mr. Rubio ’s obstinate support for the failed embargo on Cuba , which puts him on the right side of history and , increasingly , public opinion in Florida , \" it stated .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"embargo\", \"transaction.transaction.n/a\", \"An unspecified transaction involving a policy or action related to economic restrictions or trade policies.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIt is , in other words , one reason that people in these communities may feel the need to take the law into their own hands — sometimes leading to more violent crime .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"violent\", \"life.injure.n/a\", \"This event represents a personal injury incident where individuals suffer harm, often due to conflict or lack of legal recourse in communities.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThe documents were believed to be stolen by Russian insiders and then made their way to U.S. dealers and auction houses .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"auction\", \"transaction.transferownership.n/a\", \"Represents the transfer of ownership of documents from one party to another through auction mechanisms.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nWhile Mr. Bush has assailed Mr. Trump in the past and described him as all but unqualified for the presidency , the Saturday debate was his most forceful performance of the race .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"assailed\", \"conflict.attack.n/a\", \"An attack event where one party criticizes or opposes another, indicating hostility or disapproval in a political or social context.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nThese \" New Democrats \" sold themselves as centrist reformers but behaved more like merchant bankers and , within a few years , ushered in a raft of privatizations , Wall Street giveaways , tough - on - crime laws , and deregulated trade rules that would rob the middle class and set a course toward financial ruin .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"rob\", \"conflict.attack.n/a\", \"This event denotes an act of taking property or resources from someone through force or deception, often involving economic harm.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nKirk , Marco Rubio Marco RubioOvernight Finance : Ryan seeks manufacturing muscle for tax reform | Warren targets Wells Fargo board | Senators raise concerns over Russian takeover of Citgo | Pelosi hits GOP for budget delays Senators ask administration to probe possible Russian takeover of Citgo Week ahead : Obama DHS chief to testify on Russian meddling MORE ( Fla. ) , Kelly Ayotte Kelly AyotteKelly Ayotte joins defense contractor 's board of directors Week ahead : Comey firing dominates Washington GOP senators pitch Merrick Garland for FBI director MORE ( N.H. ) , David Vitter David VitterOvernight Energy : Trump set to propose sharp cuts to EPA , energy spending Former La.\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"testify\", \"justice.initiatejudicialprocess.n/a\", \"An official begins a legal or judicial process, often involving testimony or investigation in a formal setting.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nA woman who has accused Bill Clinton of rape and been a vocal opponent of Hillary Clinton 's criticized Republican presidential nominee Donald Trump 's campaign on Sunday for using her words in a political attack ad in a new interview .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"rape\", \"life.injure.injurycausedbyviolentevents\", \"This event represents physical or psychological harm caused by violent actions or events involving human interaction.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nEnraged at the execution , Shi’ite protesters in Iran ransacked and firebombed the Saudi embassy .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"ransacked\", \"artifactexistence.damagedestroy.n/a\", \"This event denotes the destruction or severe damage of an object or structure through violent or unauthorized actions by individuals.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIf we opened that border , millions of Mexicans would migrate immediately because of the higher pay and better living conditions here .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"migrate\", \"movement.transportperson.grantentryasylum\", \"This event represents granting asylum to a person for entry into a country, allowing them to relocate due to safety or economic reasons.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nA Chinese official said dialogue was needed to resolve issues on the Korean peninsula .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"dialogue\", \"contact.discussion.meet\", \"A formal meeting or discussion between parties to address and resolve issues concerning the Korean peninsula.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nObama ’s reticence has robbed Secretary of State John Kerry , who is valiantly trying to negotiate a lasting Syrian cease - fire , of leverage .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"robbed\", \"conflict.attack.stealrobhijack\", \"An entity loses control or advantage due to an act of theft or force in a conflict context, reducing strategic power or influence.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nViktor Zolotov , ex - commander of the Internal Troops and former head of the President 's personal security service , has been appointed as the leader of the new structure , with orders to report directly to the president , Kremlin spokesman Dmitry Peskov told journalists on Tuesday .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"appointed\", \"personnel.startposition.hiring\", \"A person is formally assigned to a new leadership role within an organization or government structure.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nStunningly , this SC report reveals that the reason for the CIA ordering all of their aircraft supporting their coup to be grounded was due to President Putin ’s “ threat / promise ” to immediately activate the MoD ’s S-500 Prometey ( 55R6 M Triumfator - M ) air defense systems stationed at Khmeimim Air Base that would , within 120 seconds , down every “ hostile ” aircraft over Turkey , Syria and Iraq — as well as destroy every single US - NATO military satellite in low Earth orbit ( LEO ) over the Levant .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"coup\", \"conflict.attack.n/a\", \"An unspecified conflict attack event indicating a potential military or political intervention without defined attack parameters or targets.\"]]}"} {"system_prompt": "You are an open-domain event extraction system. Your task is to identify events expressed or clearly implied in a given text.\nIMPORTANT:Output ONLY valid JSON. No explanations, no markdown, no extra text.\nOutput Format (JSON only, no markdown):\n\n{\"events\": [[, , ], [, , ]]} \n\n- If no events are detected, return: {\"events\": []}", "user_prompt": "Given an input text: \n\nIt comes as Russia vowed to shoot down any American fighter planes that attack President Bashar al - Assad 's forces in Syria .\n\n\nYour task is to extract all events present in the text. The text may contain zero, one, or multiple events.\n\nFor each event:\n- Identify a trigger: the word or phrase that most clearly indicates the event.\n- Identify event type\n- Description: explaining what this event type means\n\nConstraints and Guidelines\n- Do not invent information not supported by the text.\n- Do not paraphrase triggers or arguments.\n- The \"text\" field must exactly match a span in the original input (from ...).\n", "response": "{\"events\": [[\"vowed\", \"contact.commitmentpromiseexpressintent.meet\", \"An entity commits to take action if a specific condition is met, expressing intent to respond to an attack on Syrian forces.\"]]}"}