Research map / Large language models / Retrieval and knowledge
PIQA: Reasoning about Physical Commonsense in Natural Language
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use RoBERTa instead of GPT | Pretrained models fine-tuned for PIQA · PIQA test | accuracy: 69.2 -> 77.1 | Helped |
| Use RoBERTa instead of BERT | Pretrained models fine-tuned for PIQA · PIQA test | accuracy: 66.8 -> 77.1 | Helped |
| Rely on a large pretrained language model rather than human judgment | PIQA question answering · PIQA development | accuracy: 77.1; 94.9 | Hurt |
| Evaluate examples pivoting on water rather than spoon | RoBERTa · PIQA validation single-word edits | accuracy: 90% -> 75% | Hurt |
Closest papers by meaning
- Evaluating Commonsense in Pre-Trained Language Models (AAAI 2020 · similarity 0.88)
- Benchmarking Knowledge-Enhanced Commonsense Question Answering via Knowledge-to-Text Transformation (AAAI 2021 · similarity 0.87)
- Global PIQA: Evaluating Commonsense Reasoning Across 100+ Languages and Cultures (NeurIPS 2026 · similarity 0.86)
- Zero-Shot Commonsense Question Answering with Cloze Translation and Consistency Optimization (AAAI 2022 · similarity 0.86)
- OpenEQA: Embodied Question Answering in the Era of Foundation Models (CVPR 2024 · similarity 0.85)
- CORECODE: A Common Sense Annotated Dialogue Dataset with Benchmark Tasks for Chinese Large Language Models (AAAI 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.