Research map / Large language models / Agents, code and math
Hide and Seek with LLMs: An Adversarial Game for Sneaky Error Generation and Self-Improving Diagnosis
Rui Zou, Mengqi Wei, Yutao Zhu, Jirong Wen, Xin Zhao, Jing Chen
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 |
|---|---|---|---|
| replace each model's own diagnosis with HSG-trained diagnosis | correction by Qwen3-4B/8B/14B, DeepSeek V3, GPT-4o · GSM8K, MATH, NuminaMath-TIR | correction accuracy (average): 42.44% -> 66.15% (Qwen3-4B); 35.61% -> 66.96% (GPT-4o) | Improved |
| add adversarial rater signal to RL for the error generator (LLM-rater Adv+RL vs RL only) (ablation) | Qwen3-4B sneaky generator · NuminaMath-TIR | correction failure rate: 27.43% -> 38.51% | Improved |
| use HSG dialogue-based adversarial game instead of LLM-rater adversarial RL (ablation) | Qwen3-4B sneaky generator · NuminaMath-TIR | correction failure rate: 38.51% -> 56.17% | Improved |
| HSG dialogue game removes trivial Type A/B errors | Qwen3-4B sneaky generator · NuminaMath-TIR | share of errors with neither type: 92.6% neither type; LLM-rater Adv+RL 94.6% Type B; RL only 31.9% Type A | Improved |
| scale diagnosed model up (Qwen3-4B to 14B) with HSG diagnosis (ablation) | gain from D* · Average of three test sets | correction accuracy gain: +23.71 -> +20.06 | Mixed |
Closest papers by meaning
- SPC: Evolving Self-Play Critic via Adversarial Games for LLM Reasoning (NeurIPS 2025 · similarity 0.87)
- Seirênes: Adversarial Self-Play with Evolving Distractions for LLM Reasoning (NeurIPS 2026 · similarity 0.86)
- Generative Adversarial Reasoner: Enhancing LLM Reasoning with Adversarial Reinforcement Learning (ICLR 2026 · similarity 0.86)
- Improving Rationality in the Reasoning Process of Language Models through Self-playing Game (ICML 2025 · similarity 0.86)
- DECEIVE-AFC: Adversarial Claim Attacks against Search-Enabled LLM-based Fact-Checking Systems (NeurIPS 2026 · similarity 0.85)
- Learning Robust Reasoning through Guided Adversarial Self-Play (NeurIPS 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.
