Dataset Viewer
Auto-converted to Parquet Duplicate
id
string
pair_id
string
source_split
string
story_id
string
turn_id
int64
source
string
condition
string
is_underspecified
bool
placement
string
question
string
input_story
string
original_story
string
damaged_story
string
gold_answer_original
string
rationale_start
int64
rationale_end
int64
rationale_text
string
rationale_sentences
list
removed_sentences
list
insertion_offset
int64
menu_json
string
menu_request_json
string
menu_response_text
string
menu_contract_errors
list
recovery_kind
string
recovery_history_json
string
clarifications
list
dev:3ve8ayvf8mx6kfmvw6qjlcy49lyf8o:7:undamaged
dev:3ve8ayvf8mx6kfmvw6qjlcy49lyf8o:7
dev
3ve8ayvf8mx6kfmvw6qjlcy49lyf8o
7
gutenberg
undamaged
false
appended
Why didn't Peter think the structure was a home?
CHAPTER XI Peter Rabbit and Jerry Muskrat Are Puzzled. Jerry Muskrat was more and more sure that his big cousin, Paddy the Beaver, didn't know quite so much as he might about house-building. Jerry would have liked to offer some suggestions, but he didn't quite dare. You see, he was very anxious not to displease his b...
CHAPTER XI Peter Rabbit and Jerry Muskrat Are Puzzled. Jerry Muskrat was more and more sure that his big cousin, Paddy the Beaver, didn't know quite so much as he might about house-building. Jerry would have liked to offer some suggestions, but he didn't quite dare. You see, he was very anxious not to displease his b...
CHAPTER XI Peter Rabbit and Jerry Muskrat Are Puzzled. Jerry Muskrat was more and more sure that his big cousin, Paddy the Beaver, didn't know quite so much as he might about house-building. Jerry would have liked to offer some suggestions, but he didn't quite dare. You see, he was very anxious not to displease his b...
It doesn't look it
1,038
1,196
"I'm not stupid!" retorted Peter. "How was I to know that that pile of mud and sticks is meant for a house? It certainly doesn't look it. Where is the door?"
[ { "sentence_index": 15, "start": 1038, "end": 1055, "text": "\"I'm not stupid!\"" }, { "sentence_index": 16, "start": 1056, "end": 1073, "text": "retorted Peter. \"" }, { "sentence_index": 17, "start": 1073, "end": 1145, "text": "How was I to know that that pi...
[]
1,840
{"constraints": [], "gap": "", "items": [{"exclusions": [], "id": "unchanged", "probability": 1, "scenario": "The visible story already supports an answer; append nothing."}], "status": "already_answerable"}
{"input": "Input JSON:\n{\"story\": \"CHAPTER XI Peter Rabbit and Jerry Muskrat Are Puzzled. \\n\\nJerry Muskrat was more and more sure that his big cousin, Paddy the Beaver, didn't know quite so much as he might about house-building. Jerry would have liked to offer some suggestions, but he didn't quite dare. You see, ...
{"status":"already_answerable","gap":"","constraints":[],"items":[{"id":"unchanged","scenario":"The visible story already supports an answer; append nothing.","probability":1,"exclusions":[]}]}
[]
null
null
[ { "sample_index": 0, "status": "already_answerable", "addition": "", "applied_addition": "", "justification": "The story says Peter did not think it was a home because it looked like a pile of mud and sticks rather than a house.", "clarified_story": "CHAPTER XI Peter Rabbit and Jerry Muskrat...
dev:3p529iw9kyl1zm6eqvznqhkauvzflx:11:undamaged
dev:3p529iw9kyl1zm6eqvznqhkauvzflx:11
dev
3p529iw9kyl1zm6eqvznqhkauvzflx
11
gutenberg
undamaged
false
appended
What body was under Gonzalo de Cordoba's command?
CHAPTER III. THE ROMAN BARONS Having driven Charles VIII out of Italy, it still remained for the allies to remove all traces of his passage from Naples and to restore the rule of the House of Aragon. In this they had the aid of Ferdinand and Isabella, who sent an army under the command of that distinguished soldier G...
CHAPTER III. THE ROMAN BARONS Having driven Charles VIII out of Italy, it still remained for the allies to remove all traces of his passage from Naples and to restore the rule of the House of Aragon. In this they had the aid of Ferdinand and Isabella, who sent an army under the command of that distinguished soldier G...
CHAPTER III. THE ROMAN BARONS Having driven Charles VIII out of Italy, it still remained for the allies to remove all traces of his passage from Naples and to restore the rule of the House of Aragon. He landed in Calabria in the spring of 1496, and war broke out afresh through that already sorely devastated land. ...
an army
263
337
an army under the command of that distinguished soldier Gonzalo de Cordoba
[ { "sentence_index": 2, "start": 202, "end": 377, "text": "In this they had the aid of Ferdinand and Isabella, who sent an army under the command of that distinguished soldier Gonzalo de Cordoba, known in his day as the Great Captain." } ]
[]
1,513
{"constraints": [], "gap": "", "items": [{"exclusions": [], "id": "unchanged", "probability": 1, "scenario": "The visible story already supports an answer; append nothing."}], "status": "already_answerable"}
{"input": "Input JSON:\n{\"story\": \"CHAPTER III. THE ROMAN BARONS \\n\\nHaving driven Charles VIII out of Italy, it still remained for the allies to remove all traces of his passage from Naples and to restore the rule of the House of Aragon. In this they had the aid of Ferdinand and Isabella, who sent an army under t...
{"status":"already_answerable","gap":"","constraints":[],"items":[{"id":"unchanged","scenario":"The visible story already supports an answer; append nothing.","probability":1,"exclusions":[]}]}
[]
null
null
[ { "sample_index": 0, "status": "already_answerable", "addition": "", "applied_addition": "", "justification": "The story states that Gonzalo de Cordoba commanded an army sent by Ferdinand and Isabella.", "clarified_story": "CHAPTER III. THE ROMAN BARONS \n\nHaving driven Charles VIII out of ...
dev:3ftop5warfo47s3oks4p7vkek7uj0q:13:undamaged
dev:3ftop5warfo47s3oks4p7vkek7uj0q:13
dev
3ftop5warfo47s3oks4p7vkek7uj0q
13
wikipedia
undamaged
false
appended
What is a synonym of endemic?
Endemism is the ecological state of a species being unique to a defined geographic location, such as an island, nation, country or other defined zone, or habitat type; organisms that are indigenous to a place are not endemic to it if they are also found elsewhere. The extreme opposite of endemism is cosmopolitan distri...
Endemism is the ecological state of a species being unique to a defined geographic location, such as an island, nation, country or other defined zone, or habitat type; organisms that are indigenous to a place are not endemic to it if they are also found elsewhere. The extreme opposite of endemism is cosmopolitan distri...
Endemism is the ecological state of a species being unique to a defined geographic location, such as an island, nation, country or other defined zone, or habitat type; organisms that are indigenous to a place are not endemic to it if they are also found elsewhere. The extreme opposite of endemism is cosmopolitan distri...
Precinctive
669
818
The term "precinctive" has been suggested by some scientists, and was first used in botany by MacCaughey in 1917. It is the equivalent of "endemism".
[ { "sentence_index": 5, "start": 669, "end": 782, "text": "The term \"precinctive\" has been suggested by some scientists, and was first used in botany by MacCaughey in 1917." }, { "sentence_index": 6, "start": 783, "end": 820, "text": "It is the equivalent of \"endemism\". \"" ...
[]
1,632
{"constraints": [], "gap": "", "items": [{"exclusions": [], "id": "unchanged", "probability": 1, "scenario": "The visible story already supports an answer; append nothing."}], "status": "already_answerable"}
{"input": "Input JSON:\n{\"story\": \"Endemism is the ecological state of a species being unique to a defined geographic location, such as an island, nation, country or other defined zone, or habitat type; organisms that are indigenous to a place are not endemic to it if they are also found elsewhere. The extreme oppos...
{"status":"already_answerable","gap":"","constraints":[],"items":[{"id":"unchanged","scenario":"The visible story already supports an answer; append nothing.","probability":1,"exclusions":[]}]}
[]
null
null
[ { "sample_index": 0, "status": "already_answerable", "addition": "", "applied_addition": "", "justification": "The story explicitly identifies “precinctive” as an alternative term for endemic.", "clarified_story": "Endemism is the ecological state of a species being unique to a defined geogr...
dev:317hq483i7sbxdbp3gln661reeqnio:4:undamaged
dev:317hq483i7sbxdbp3gln661reeqnio:4
dev
317hq483i7sbxdbp3gln661reeqnio
4
mctest
undamaged
false
appended
Who were the receipients of Mrs. Smith's pies?
"Mrs. Smith liked to deliver her pies with her big hot air balloon. \n\nMrs. Smith like baking pies (...TRUNCATED)
"Mrs. Smith liked to deliver her pies with her big hot air balloon. \n\nMrs. Smith like baking pies (...TRUNCATED)
"Mrs. Smith liked to deliver her pies with her big hot air balloon. \n\nMrs. Smith like baking pies (...TRUNCATED)
her neighbors
121
163
She baked for her neighbors of all ages.
[ { "sentence_index": 2, "start": 122, "end": 162, "text": "She baked for her neighbors of all ages." } ]
[]
1,058
"{\"constraints\": [], \"gap\": \"\", \"items\": [{\"exclusions\": [], \"id\": \"unchanged\", \"prob(...TRUNCATED)
"{\"input\": \"Input JSON:\\n{\\\"story\\\": \\\"Mrs. Smith liked to deliver her pies with her big h(...TRUNCATED)
"{\"status\":\"already_answerable\",\"gap\":\"\",\"constraints\":[],\"items\":[{\"id\":\"unchanged\"(...TRUNCATED)
[]
null
null
[{"sample_index":0,"status":"already_answerable","addition":"","applied_addition":"","justification"(...TRUNCATED)
dev:3jv9lgbjwtefj756e7lx0jogp3ygoj:17:undamaged
dev:3jv9lgbjwtefj756e7lx0jogp3ygoj:17
dev
3jv9lgbjwtefj756e7lx0jogp3ygoj
17
cnn
undamaged
false
appended
What would be Paul's job in Russia?
"(CNN) -- A Russian bookmaking company is offering to pay 100,000 euros for Paul, the prognosticatin(...TRUNCATED)
"(CNN) -- A Russian bookmaking company is offering to pay 100,000 euros for Paul, the prognosticatin(...TRUNCATED)
"(CNN) -- A Russian bookmaking company is offering to pay 100,000 euros for Paul, the prognosticatin(...TRUNCATED)
to forecast football games, and act as a mascot fot the bookmaker
725
1,140
"\"Seriously speaking, we want the octopus for a number of purposes,\" Zhuravsky said. \"First, to s(...TRUNCATED)
[{"sentence_index":6,"start":725,"end":810,"text":"\"Seriously speaking, we want the octopus for a n(...TRUNCATED)
[]
1,586
"{\"constraints\": [], \"gap\": \"\", \"items\": [{\"exclusions\": [], \"id\": \"unchanged\", \"prob(...TRUNCATED)
"{\"input\": \"Input JSON:\\n{\\\"story\\\": \\\"(CNN) -- A Russian bookmaking company is offering t(...TRUNCATED)
"{\"status\":\"already_answerable\",\"gap\":\"\",\"constraints\":[],\"items\":[{\"id\":\"unchanged\"(...TRUNCATED)
[]
null
null
[{"sample_index":0,"status":"already_answerable","addition":"","applied_addition":"","justification"(...TRUNCATED)
dev:3ymtujh0dsgfkjhufn5vl4x0y7f4th:1:undamaged
dev:3ymtujh0dsgfkjhufn5vl4x0y7f4th:1
dev
3ymtujh0dsgfkjhufn5vl4x0y7f4th
1
gutenberg
undamaged
false
appended
How did Guy feel in the morning?
"CHAPTER 33 \n\nFrom darkness here and dreariness, We ask not full repose. --CHRISTIAN YEAR \n\nIt s(...TRUNCATED)
"CHAPTER 33 \n\nFrom darkness here and dreariness, We ask not full repose. --CHRISTIAN YEAR \n\nIt s(...TRUNCATED)
"CHAPTER 33 \n\nFrom darkness here and dreariness, We ask not full repose. Both he and Amabel sat f(...TRUNCATED)
dull and weary with a headache
186
254
he had a slight headache the next morning, and seemed dull and weary
[{"sentence_index":1,"start":72,"end":255,"text":"--CHRISTIAN YEAR \n\nIt seemed as if the fatigue w(...TRUNCATED)
[]
1,436
"{\"constraints\": [], \"gap\": \"\", \"items\": [{\"exclusions\": [], \"id\": \"unchanged\", \"prob(...TRUNCATED)
"{\"input\": \"Input JSON:\\n{\\\"story\\\": \\\"CHAPTER 33 \\\\n\\\\nFrom darkness here and drearin(...TRUNCATED)
"{\"status\":\"already_answerable\",\"gap\":\"\",\"constraints\":[],\"items\":[{\"id\":\"unchanged\"(...TRUNCATED)
[]
null
null
[{"sample_index":0,"status":"already_answerable","addition":"","applied_addition":"","justification"(...TRUNCATED)
dev:340ugxu9dy1te9fyzowszqjfvbsuvw:17:undamaged
dev:340ugxu9dy1te9fyzowszqjfvbsuvw:17
dev
340ugxu9dy1te9fyzowszqjfvbsuvw
17
race
undamaged
false
appended
How did the police thank Vicki Brown?
"One of Britain's bravest women told yesterday how she helped to catch suspected police killer David(...TRUNCATED)
"One of Britain's bravest women told yesterday how she helped to catch suspected police killer David(...TRUNCATED)
" It was also said that she could _ a share of up to $30,000 reward money. \n\nVicki Brown, 30, play(...TRUNCATED)
with flowers
112
150
was thanked with flowers by the police
[{"sentence_index":0,"start":0,"end":151,"text":"One of Britain's bravest women told yesterday how s(...TRUNCATED)
[]
1,973
"{\"constraints\": [], \"gap\": \"\", \"items\": [{\"exclusions\": [], \"id\": \"unchanged\", \"prob(...TRUNCATED)
"{\"input\": \"Input JSON:\\n{\\\"story\\\": \\\"One of Britain's bravest women told yesterday how s(...TRUNCATED)
"{\"status\":\"already_answerable\",\"gap\":\"\",\"constraints\":[],\"items\":[{\"id\":\"unchanged\"(...TRUNCATED)
[]
null
null
[{"sample_index":0,"status":"already_answerable","addition":"","applied_addition":"","justification"(...TRUNCATED)
dev:3r6byfzzp7cwzgn34e2b1bfx1y9xfp:20:undamaged
dev:3r6byfzzp7cwzgn34e2b1bfx1y9xfp:20
dev
3r6byfzzp7cwzgn34e2b1bfx1y9xfp
20
race
undamaged
false
appended
Was ChiChi successful in saving the women?
"ChiChi weighs only 13 pounds. \"He's so tiny,I can carry him with one hand,\" says Mary Lane.\"Most(...TRUNCATED)
"ChiChi weighs only 13 pounds. \"He's so tiny,I can carry him with one hand,\" says Mary Lane.\"Most(...TRUNCATED)
"ChiChi weighs only 13 pounds. \"He's so tiny,I can carry him with one hand,\" says Mary Lane.\"Most(...TRUNCATED)
yes
1,786
1,833
.They thanked the Lanes for saving their lives.
[{"sentence_index":26,"start":1745,"end":1787,"text":"Luckily,the women were fine,though shaken."},{(...TRUNCATED)
[]
1,942
"{\"constraints\": [], \"gap\": \"\", \"items\": [{\"exclusions\": [], \"id\": \"unchanged\", \"prob(...TRUNCATED)
"{\"input\": \"Input JSON:\\n{\\\"story\\\": \\\"ChiChi weighs only 13 pounds. \\\\\\\"He's so tiny,(...TRUNCATED)
"{\"status\":\"already_answerable\",\"gap\":\"\",\"constraints\":[],\"items\":[{\"id\":\"unchanged\"(...TRUNCATED)
[]
null
null
[{"sample_index":0,"status":"already_answerable","addition":"","applied_addition":"","justification"(...TRUNCATED)
dev:3z3zlgnnsiuha76yy56h6uu712v3qx:3:undamaged
dev:3z3zlgnnsiuha76yy56h6uu712v3qx:3
dev
3z3zlgnnsiuha76yy56h6uu712v3qx
3
wikipedia
undamaged
false
appended
What country is Plymouth located in?
"Plymouth (i/ˈplɪməθ/) is a city on the south coast of Devon, England, about 37 miles (60 km) so(...TRUNCATED)
"Plymouth (i/ˈplɪməθ/) is a city on the south coast of Devon, England, about 37 miles (60 km) so(...TRUNCATED)
" \n\nPlymouth's early history extends to the Bronze Age, when a first settlement emerged at Mount B(...TRUNCATED)
England
61
68
England
[{"sentence_index":0,"start":0,"end":302,"text":"Plymouth (i/ˈplɪməθ/) is a city on the south co(...TRUNCATED)
[]
1,749
"{\"constraints\": [], \"gap\": \"\", \"items\": [{\"exclusions\": [], \"id\": \"unchanged\", \"prob(...TRUNCATED)
"{\"input\": \"Input JSON:\\n{\\\"story\\\": \\\"Plymouth (i/ˈplɪməθ/) is a city on the south co(...TRUNCATED)
"{\"status\":\"already_answerable\",\"gap\":\"\",\"constraints\":[],\"items\":[{\"id\":\"unchanged\"(...TRUNCATED)
[]
null
null
[{"sample_index":0,"status":"already_answerable","addition":"","applied_addition":"","justification"(...TRUNCATED)
dev:36zn444ytrytfyb14vl0lv1w5udiot:9:undamaged
dev:36zn444ytrytfyb14vl0lv1w5udiot:9
dev
36zn444ytrytfyb14vl0lv1w5udiot
9
wikipedia
undamaged
false
appended
What did the creation of the National Assembly for Wales follow?
"The National Assembly for Wales (; commonly known as the Welsh Assembly) is a devolved assembly wit(...TRUNCATED)
"The National Assembly for Wales (; commonly known as the Welsh Assembly) is a devolved assembly wit(...TRUNCATED)
"The National Assembly for Wales (; commonly known as the Welsh Assembly) is a devolved assembly wit(...TRUNCATED)
a referendum
569
607
Act 1998, which followed a referendum
[{"sentence_index":3,"start":517,"end":615,"text":"The Assembly was created by the Government of Wal(...TRUNCATED)
[]
2,143
"{\"constraints\": [], \"gap\": \"\", \"items\": [{\"exclusions\": [], \"id\": \"unchanged\", \"prob(...TRUNCATED)
"{\"input\": \"Input JSON:\\n{\\\"story\\\": \\\"The National Assembly for Wales (; commonly known a(...TRUNCATED)
"{\"status\":\"already_answerable\",\"gap\":\"\",\"constraints\":[],\"items\":[{\"id\":\"unchanged\"(...TRUNCATED)
[]
null
null
[{"sample_index":0,"status":"already_answerable","addition":"","applied_addition":"","justification"(...TRUNCATED)
End of preview. Expand in Data Studio

CoQAR Clarifications

This dataset pairs 1,000 CoQAR development questions with their original stories and stories damaged by sentence deletion. Each of the resulting 2,000 inputs has five sampled model clarifications. Two configurations reuse the same generated additions and differ only in where those additions are placed.

Configuration Rows in dev Clarifications per row Placement
appended 2,000 5 At the end of the input story
inserted 2,000 5 At the deleted passage location for damaged inputs; after the corresponding rationale sentences for undamaged inputs

Each configuration contains 1,000 damaged and 1,000 undamaged rows. The two configurations contain the same 10,000 sampled additions, rendered in different positions; they are not separate generation runs.

from datasets import load_dataset

appended = load_dataset("rvashurin/coqar-clarifications", "appended", split="dev")
inserted = load_dataset("rvashurin/coqar-clarifications", "inserted", split="dev")

Construction

The source is the CoQAR development JSON from Orange-OpenSource/COQAR, pinned to revision 1f21713f6d3152232bb1921a3767d2b7fa7255b4:

  • Source file: data/CoQAR/dev/coqar-dev-v1.0.json.
  • Source SHA-256: 3d093e1bf45e44f7cc4e3b8e07a37ded9162ded4ac88e90bb9e4c5657be3251f.
  • Subset selection seed: 20260911.
  • Selection: 1,000 distinct eligible question turns, covering 437 stories, after excluding the ten turns used in the earlier pilot. This is a turn-level exclusion; other turns from the same stories may be present.
  • Question: the first nonempty CoQAR out-of-context rewrite for each selected turn.
  • Damage: remove every complete sentence that overlaps the original annotated answer rationale. Sentence spans and rationale spans use the original story's character coordinates. Surviving characters, including whitespace between deleted sentences, are preserved.

The row flag is_underspecified is exactly condition == "damaged". It records nominal underspecification induced by the construction, not a judgment that the question is unanswerable. The damaged story can still contain enough information to answer the question. Likewise, undamaged does not guarantee that every rewritten question is well specified or correctly answerable.

Clarification generation

The two-step v2 pipeline used gpt-5.6-luna with reasoning effort low and temperature 1. First, the model produced a menu of possible interpretations, including an already-answerable state when appropriate. Five interpretation draws were then made with replacement from the normalized model-provided weights, using reproducible draws derived from seed 20260911. Each selected interpretation was passed to a separate clarification-generation call. Duplicates and empty additions were retained.

The model saw the input story and the question; the rendering step also received the interpretation menu and selected interpretation. It did not receive the original gold answer, deleted sentences, or rationale offsets. Those offsets are used later to construct the inserted configuration.

Five draws share the first-stage menu. In particular, an already-answerable menu can produce five empty additions; these are not five independent answerability judgments. The menu weights and generated additions are model outputs, not calibrated estimates of a human clarification distribution.

Four structural failures were recovered with their initial failures and recovery histories preserved. Outputs were not filtered or resampled for semantic quality, novelty, or diversity. Two menu errors involving reserved interpretation IDs caused three damaged-input empty additions; the affected records remain present with contract errors.

Placement rules

In appended, a nonempty trimmed addition is applied as input_story + "\n\n" + applied_addition. An empty addition leaves the input story exactly unchanged. This reproduces the original generation run's story construction.

In inserted, the same trimmed addition is inserted:

  • For a damaged input, at the start of the first removed sentence, expressed in the damaged input's coordinates.
  • For an undamaged input, immediately after the last original sentence that overlaps the rationale span. There is no random placement.

A nonempty addition is separated from each nonempty side of the input by a paragraph separator ("\n\n"). Every existing input character is preserved, including whitespace at the insertion boundary. An empty addition leaves the input exactly unchanged. Placement moves the addition; it does not regenerate its wording to fit the new location.

Fields

Each row is one context–question input and contains the following fields:

Field Meaning
id, pair_id Row identity and identity shared by the damaged/undamaged pair
condition, is_underspecified damaged or undamaged, and its nominal damage flag
placement appended or inserted
question CoQAR question presented to the model
input_story Story presented to the model, before adding a clarification
original_story, damaged_story Original source story and its sentence-deleted counterpart
gold_answer_original Original CoQA answer, retained for reference
source, story_id, turn_id Original domain/source and identifiers
rationale_start, rationale_end, rationale_text Original annotated rationale and its source-story coordinates
rationale_sentences Original sentence spans overlapping the rationale
removed_sentences Sentences actually deleted for this input; empty for undamaged rows
insertion_offset Position in input_story at which the clarification block is placed
menu_json, menu_request_json, menu_response_text Interpretation menu, exact API request, and model response text
menu_contract_errors Recorded menu contract violations
recovery_kind, recovery_history_json Structural recovery information where applicable
clarifications A list of five sample records, described below

Each element of clarifications contains:

Field Meaning
sample_index, status Sample identity and generated status
addition, applied_addition, justification Raw model addition, its trimmed applied form, and model justification
clarified_story, unchanged Rendered story and whether it equals input_story exactly
selected_id, uniform_draw, seed_text, normalization Interpretation selection and reproducibility data
inserted_block_start, inserted_block_end Boundaries of the entire inserted block, including added separators
addition_start, addition_end Boundaries of the applied addition within clarified_story
prefix_separator, suffix_separator Separators added around the clarification
generation_request_json, generation_response_text Exact request and response text from the original clarification call
audit_json, audit_applies_to_placement Retained automated assessment and the placement to which it applies
contract_errors Recorded output contract violations

Character offsets are zero-based; span ends are exclusive. They index Unicode characters, not encoded bytes. Original rationale and sentence coordinates refer to original_story; clarification and inserted-block coordinates refer to clarified_story. Request and response fields describe the original generation calls, including when those additions are later placed in inserted stories.

Assessment and limitations

Damaged outputs were assessed by Luna with reasoning effort none, using the appended stories. These assessments are automated judgments, not human validation. Undamaged outputs were not audited by that step. The inserted stories have not been re-audited; any retained audit pertains to the appended version, as indicated by audit_applies_to_placement.

There are 4,550 unchanged undamaged samples and 1,608 unchanged damaged samples in each configuration. In the damaged condition, 321 inputs contribute five unchanged samples each; the remaining three unchanged samples are the recorded menu-ID errors. Empty additions are preserved in both placements.

The original gold answer is not a target label for a hypothetical clarified story: a sampled addition may establish a different answer. Model additions may be incorrect, contradictory, or unnecessary. Moving them into a passage can also affect coherence. The damage flag, model menu, audits, and source gold answer represent different kinds of information and should not be treated as interchangeable ground truth.

Original stories, rationale annotations, and original answers are exposed for analysis. An evaluation intended to measure behavior on damaged inputs should keep those reference fields out of model inputs.

Reproduction and provenance

The bundled code/export_hub.py constructs both configurations from the original saved experiment artifacts without further model calls. code/publish_hub.py publishes the prepared release. The original large validation JSONs are local run artifacts, identified by hashes in provenance.json; they are not included in this Hub repository. The published rows retain the input text, raw additions, exact requests/responses and insertion offsets needed to reconstruct either placement directly. The bundled processing sources and prompts document seeded selection, sentence removal, interpretation sampling, API processing, structural recovery, and export. Source and artifact hashes trace the published records back to the run. Public provenance omits local absolute paths and host identity.

Licensing and attribution

The CoQAR authors' license description assigns CC BY-SA 4.0 to their annotations and Apache 2.0 to their code. The original CoQA passages retain their domain-specific terms: CC BY-SA 4.0 for literature/Wikipedia, MSR-LA for MCTest, RACE-specific terms for examination passages, and Apache terms for CNN material. See the official CoQA license description and this repository's LICENSE.

This release is a derivative research dataset. Its sentence deletions, model-generated additions, and alternative placements are changes to the source material. Publishing these additions does not relicense the underlying passages. The license metadata is other to reflect the inherited mixture of terms; it is not a declaration that all components share one permissive license.

Citations

Please cite the source datasets:

@inproceedings{brabant-etal-2022-coqar,
  title = "{C}o{QAR}: Question Rewriting on {C}o{QA}",
  author = "Brabant, Quentin and Lecorv{\'e}, Gw{\'e}nol{\'e} and Rojas Barahona, Lina M.",
  booktitle = "Proceedings of the Thirteenth Language Resources and Evaluation Conference",
  year = "2022",
  pages = "119--126",
  publisher = "European Language Resources Association",
  url = "https://aclanthology.org/2022.lrec-1.13/"
}

@article{reddy-etal-2019-coqa,
  title = "{C}o{QA}: A Conversational Question Answering Challenge",
  author = "Reddy, Siva and Chen, Danqi and Manning, Christopher D.",
  journal = "Transactions of the Association for Computational Linguistics",
  volume = "7",
  year = "2019",
  pages = "249--266",
  doi = "10.1162/tacl_a_00266",
  url = "https://aclanthology.org/Q19-1016/"
}
Downloads last month
77