Update README.md

README.md

@@ -1,3 +1,95 @@
----
-license: apache-2.0
----
+---
+license: cc-by-4.0
+task_categories:
+  - question-answering
+  - multiple-choice
+language:
+  - en
+tags:
+  - multi-choice
+  - question-answering
+  - logical-reasoning
+  - compositional-reasoning
+  - commonsense-reasoning
+pretty_name: LOGICAL-COMMONSENSEQA
+size_categories:
+  - 10K<n<100K
+---
+
+## LOGICAL-COMMONSENSEQA
+
+LOGICAL-COMMONSENSEQA reframes commonsense reasoning as logical composition over pairs of atomic statements using plausibility-level operators, AND, OR, and NEITHER/NOR. Each instance has a commonsense question and four candidate answer options, and every option joins two atomic answers under one of these operators. Exactly one option is valid per instance.
+
+Most commonsense benchmarks rely on single-label evaluation, so it's never clear whether a model actually judges the individual statements correctly or just gets lucky on the final label. This benchmark separates the two: a model can score every atomic statement correctly and still fail if it can't combine them under the operator.
+
+It was introduced in "LOGICAL-COMMONSENSEQA: A Benchmark for Logical Commonsense Reasoning" (Junias and Pacheco, 2026), which evaluates instruction-tuned, reasoning-specialized, and fine-tuned models under zero-shot, few-shot, and chain-of-thought prompting. Models handle conjunctive composition reasonably well and disjunctive composition moderately, but performance degrades sharply on negation-based questions. The benchmark is also used as one of two evaluation sets in "From Atomic Evidence to Logical Composition: Structured Compositional Reasoning over Compound Answer Options" (Junias and Pacheco, 2026), paired there with LOGICAL-SATA, a reading-comprehension counterpart built the same way from SATA-Bench.
+
+### Example
+
+```json
+{"question": "Sammy wanted to go to where the people were. Where might he go?", "choices": ["social venues AND empty parks", "community gatherings AND social venues", "social venues AND train platforms", "community gatherings AND quiet retreats"], "label": 1, "qa_type": "AND"}
+```
+
+Both atoms in the labeled option need to hold for it to be valid, and validity is determined jointly by the atomic statuses and the operator, not by scoring each option on its own.
+
+### Construction
+
+Each instance contains a commonsense question and four compound answer options, each joining two atomic answers under a logical operator. The full set of 19,996 instances is evenly distributed across four settings: three in which all four options in an instance share the same operator (AND, OR, NEITHER/NOR), and a MIXED setting in which the operators can differ across the options within a single instance.
+
+### What's in this release
+
+This release has the train and validation splits only. The test split used for the numbers reported in both papers is not included here, to keep it usable as a held-out evaluation set. If you need it for reproducibility or direct comparison, reach out (contact below).
+
+| split | instances |
+|---|---|
+| train | 11,996 |
+| validation | 6,000 |
+| test (not released) | 2,000 |
+
+The original test split was further divided into human-validated (HV) and non-validated (NV) subsets of 1,000 each. Both papers report primary results on the HV subset.
+
+Each split is balanced across the four operator settings, AND, OR, NEITHER/NOR, MIXED.
+
+### Fields
+
+- `question` - the commonsense question
+- `choices` - four compound answer options, each a pair of atomic answers joined by an operator, e.g. `"community gatherings AND social venues"`
+- `label` - index (0-based) of the single valid option in `choices`
+- `qa_type` - operator setting for the instance: `AND`, `OR`, `NEITHER`, or `Mixed`
+
+### License
+
+This dataset is released under CC BY 4.0. If you use this dataset, please attribute both papers below.
+
+### Citation
+
+```
+@inproceedings{junias-pacheco-2026-logical,
+    title = "{LOGICAL}-{COMMONSENSEQA}: A Benchmark for Logical Commonsense Reasoning",
+    author = "Junias, Obed  and
+      Pacheco, Maria Leonor",
+    editor = "Liakata, Maria  and
+      Moreira, Viviane P.  and
+      Zhang, Jiajun  and
+      Jurgens, David",
+    booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 2: Short Papers)",
+    month = jul,
+    year = "2026",
+    address = "San Diego, California, United States",
+    publisher = "Association for Computational Linguistics",
+    url = "https://aclanthology.org/2026.acl-short.61/",
+    doi = "10.18653/v1/2026.acl-short.61",
+    pages = "746--758",
+    ISBN = "979-8-89176-391-3"
+}
+
+@inproceedings{junias2026atomic,
+  title={From Atomic Evidence to Logical Composition: Structured Compositional Reasoning over Compound Answer Options},
+  author={Junias, Obed and Pacheco, Maria Leonor},
+  year={2026}
+}
+```
+
+### Contact
+
+obed.junias@colorado.edu