| --- |
| pretty_name: OpenJevData-140k |
| language: |
| - en |
| - zh |
| license: mit |
| size_categories: |
| - 100K<n<1M |
| task_categories: |
| - question-answering |
| tags: |
| - decision-making |
| - reasoning |
| - supervised-fine-tuning |
| - reinforcement-learning |
| configs: |
| - config_name: default |
| data_files: |
| - split: SFT |
| path: data/SFT.parquet |
| - split: RL |
| path: data/RL.parquet |
| --- |
| |
| # Dataset Card for OpenJevData-140k |
|
|
| ## Dataset Summary |
|
|
| **OpenJevData-140k** is a curated release of the data collection used to train [**OpenJev-4B**](https://huggingface.co/shenjunhao/OpenJev-4B). It contains **146,738 decision-making examples** across **19 task categories**, organized into **SFT** and **RL** splits. |
|
|
| Each example presents a state, a question, and a request-specific set of natural-language options. The data include hard answers and soft probability distributions, with candidate sets ranging from **2 to 77 options**. Languages are **English and Chinese**. |
|
|
| ## Data Splits |
|
|
| The **SFT** split contains **100,345 examples** for supervised fine-tuning: **86,023 hard** and **14,322 soft** examples. The **RL** split contains **46,393 examples** for reinforcement learning: **28,667 hard** and **17,726 soft** examples. Together they provide **146,738 examples**. |
|
|
| ## Task Distribution |
|
|
|  |
|
|
| - **Probabilistic reasoning:** Infer outcome probabilities from uncertainty, conditional information, sampling, or noisy evidence. |
| - **Deductive reasoning:** Derive conclusions from stated premises, rules, or natural-language entailment. |
| - **Commonsense context:** Use everyday knowledge to answer a question or select a plausible continuation. |
| - **Knowledge QA:** Answer factual and subject-specific questions across multiple domains. |
| - **Policy exceptions:** Apply policies while resolving exceptions, rule precedence, and conflicting conditions. |
| - **Multi-hop reasoning:** Connect facts across passages, documents, or entities to reach an answer. |
| - **Temporal / numeric:** Apply rules involving dates, durations, numeric thresholds, or time-dependent conditions. |
| - **Quantitative reasoning:** Solve numerical questions involving arithmetic, word problems, financial information, or tables. |
| - **Intent routing:** Identify the user intent and select an appropriate tool, route, or action. |
| - **Evidence validation:** Check whether evidence supports a claim or whether a response meets stated semantic and format requirements. |
| - **Reading comprehension:** Answer questions using information stated or implied in a supplied passage. |
| - **Answer preference:** Choose the response that best satisfies the request among candidate answers. |
| - **Option logic:** Jointly interpret logical statements and dependencies distributed across the candidate options. |
| - **Score quality:** Assign a rating for qualities such as helpfulness, coherence, clarity, or preference strength. |
| - **Score rubrics:** Assign a score by applying explicit criteria for completeness, coverage, compliance, or task completion. |
| - **Constraint planning:** Choose an action or plan that satisfies interacting constraints and trade-offs. |
| - **Code / database:** Reason about code behavior or select the database query that answers a request. |
| - **Emotion recognition:** Identify the emotion conveyed by a piece of text. |
| - **Score intensity:** Rate the strength of a tone or the severity of an impact on an ordered scale. |
|
|
| ## Dataset Creation |
|
|
| The source benchmarks include [ARC](https://huggingface.co/datasets/allenai/ai2_arc), [BANKING77](https://huggingface.co/datasets/PolyAI/banking77), [BoolQ](https://github.com/google-research-datasets/boolean-questions), [CodeContests](https://github.com/google-deepmind/code_contests), [CommonsenseQA](https://www.tau-nlp.sites.tau.ac.il/commonsenseqa), [ConditionalQA](https://github.com/haitian-sun/ConditionalQA), [ContractNLI](https://stanfordnlp.github.io/contract-nli/), [FinQA](https://github.com/czyssrs/FinQA), [GoEmotions](https://github.com/google-research/google-research/tree/master/goemotions), [HellaSwag](https://rowanzellers.com/hellaswag/), [HelpSteer2](https://huggingface.co/datasets/nvidia/HelpSteer2), [HoVer](https://hover-nlp.github.io/), [LogiQA 2.0](https://github.com/csitfun/LogiQA2.0), [MASSIVE](https://huggingface.co/datasets/AmazonScience/massive), [MathQA](https://math-qa.github.io/), [MMLU auxiliary data](https://huggingface.co/datasets/cais/mmlu), [MultiDoc2Dial](https://doc2dial.github.io/multidoc2dial/), [MuSiQue](https://github.com/StonyBrookNLP/musique), [OpenBookQA](https://huggingface.co/datasets/allenai/openbookqa), [OR-ShARC](https://github.com/Yifan-Gao/open_retrieval_conversational_machine_reading), [PRM800K](https://github.com/openai/prm800k), [QASC](https://huggingface.co/datasets/allenai/qasc), [RAGTruth](https://github.com/ParticleMedia/RAGTruth), [RuleTaker](https://github.com/allenai/ruletaker), [SNLI](https://nlp.stanford.edu/projects/snli/), [Spider](https://yale-lily.github.io/spider), [TAT-QA](https://github.com/NExTplusplus/TAT-QA), [TempReason](https://github.com/DAMO-NLP-SG/TempReason), [TimeQA](https://github.com/wenhuchen/Time-Sensitive-QA), [ToolBench](https://github.com/OpenBMB/ToolBench), [WANLI](https://github.com/alisawuffles/wanli), and [When2Call](https://huggingface.co/datasets/nvidia/When2Call), together with [instruction-following constraint datasets](https://huggingface.co/datasets/allenai/IF_multi_constraints_upto5) and human-annotated ratings. These sources are converted into a shared decision format and complemented by program-generated tasks and LLM-assisted synthesis. Labels come from source annotations, executable checks, exact probability calculations, or reviewed judgments, depending on the task. |
|
|
| Synthetic data generation uses the **input token-length characteristics of JevBench Public** as a reference to broaden input-length coverage. The collection includes both short, straightforward questions and longer, more demanding decision problems. The harder tasks include multi-step reasoning, interacting rules and exceptions, cross-option logic, temporal and quantitative constraints, and decisions under uncertainty. |
|
|
| ## Data Format |
|
|
| Each example contains the following fields: |
|
|
| - `state`: context, evidence, rules, or a scenario. |
| - `question`: the requested decision. |
| - `options`: an ordered list of `{id, name, criteria}` objects. Nullable `criteria` may contain essential conditions and is part of the model input. |
| - `category`: the broad task category. |
| - `task`: `choice`, `score`, or `noul`; each uses the supplied candidate options. |
| - `score_values`: numeric values aligned with the options for scoring tasks, otherwise null. |
| - `answer`: `{kind, probabilities}`, where `kind` is `hard` or `soft`. The probabilities align by position with `options` and sum to one. A hard answer is represented by a one-hot distribution. |
|
|
| ## Loading the Dataset |
|
|
| ```python |
| from datasets import load_dataset |
| |
| dataset = load_dataset("shenjunhao/OpenJevData-140k") |
| sft = dataset["SFT"] |
| rl = dataset["RL"] |
| |
| print(len(sft), len(rl)) # 100345 46393 |
| example = sft[0] |
| options = example["options"] |
| gold_probabilities = example["answer"]["probabilities"] |
| ``` |
|
|
| ## Licensing Information |
|
|
| **MIT** applies to the original contributions in this release. Adapted source material retains its upstream license and attribution requirements; see [SOURCE_NOTICES.md](SOURCE_NOTICES.md). |
|
|
