Research map / Large language models / Retrieval and knowledge
PCoKG: Personality-aware Commonsense Reasoning with Debate
Weijie Li, Zhongqing Wang, Guodong Zhou
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add MBTI personality types to the input (ablation) | PCoKGM (LLaMA3) · PCoKG test | BLEU-4: 10.16 -> 13.73 | Improved |
| add MBTI personality types to the input (ablation) | PCoKGM (MiniCPM4) · PCoKG test | BLEU-4: 10.45 -> 14.50 | Improved |
| add event/dimension selection (ablation) | PCoKGM (LLaMA3) · PCoKG test | BLEU-4: 11.25 -> 13.73 | Improved |
| add event/dimension selection (ablation) | PCoKGM (Qwen3) · PCoKG test | BLEU-4: 10.56 -> 14.08 | Improved |
| add debate refinement to data construction (ablation) | PCoKGM (LLaMA3) · PCoKG test | BLEU-4: 12.09 -> 13.73 | Improved |
| add debate refinement to data construction (ablation) | PCoKGM (Qwen3) · PCoKG test | BLEU-4: 11.00 -> 14.08 | Improved |
| add debate refinement to data construction (ablation) | PCoKGM (MiniCPM4) · PCoKG test | BLEU-4: 11.46 -> 14.50 | Improved |
| add selection and debate together (ablation) | PCoKGM (Qwen3) · PCoKG test | BLEU-4: 10.66 -> 14.08 | Improved |
| use natural-language prompts for personality instead of special tokens (PCoKGM vs COMET) | LLaMA3-1B · PCoKG test | BLEU-4: 12.58 -> 13.73 | Improved |
| use natural-language prompts for personality instead of special tokens (PCoKGM vs COMET) | Qwen3-0.6B · PCoKG test | BLEU-4: 10.09 -> 14.08 | Improved |
| fine-tune on PCoKG vs one-shot prompting large LLM | MiniCPM4-0.5B vs o4-mini · PCoKG test | BLEU-4: 5.38 -> 14.50 | Improved |
| scale up base model size (LoRA) | Qwen3/LLaMA3/MiniCPM4 · PCoKG test | BLEU-4 and ROUGE | Improved |
Closest papers by meaning
- CORECODE: A Common Sense Annotated Dialogue Dataset with Benchmark Tasks for Chinese Large Language Models (AAAI 2024 · similarity 0.89)
- Diplomat: A Dialogue Dataset for Situated PragMATic Reasoning (NeurIPS 2023 · similarity 0.87)
- ToMATO: Verbalizing the Mental States of Role-Playing LLMs for Benchmarking Theory of Mind (AAAI 2025 · similarity 0.86)
- PersonaVLM: Long-Term Personalized Multimodal LLMs (CVPR 2026 · similarity 0.86)
- Consistently Simulating Human Personas with Multi-Turn Reinforcement Learning (NeurIPS 2025 · similarity 0.85)
- MAPS: Multi-Agent Personality Shaping for Collaborative Reasoning (AAAI 2026 · similarity 0.85)
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.
