Research map / Large language models / Agents, code and math
Scaling the Scaling Logic: Agentic Meta-Synthesis of Logic Reasoning
Bowen LIU, Zhi Wu, Runquan Xie, Zhanhui Kang, Jia Li
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 |
|---|---|---|---|
| train on agent-evolved task families instead of seed families (ablation) | GRPO on Qwen3-8B-Base · SynLogic | pass@1: 13.1 -> 18.7 (step 240) | Improved |
| train on evolved families instead of seed families (ablation) | GRPO on Qwen3-8B-Base · AIME25 | pass@1: 11.0 -> 14.0 (step 240) | Improved |
| evolve Enigmata families with the pipeline at matched instance count and steps (ablation) | Enigmata data, Qwen3-8B-Base · BBH | pass@1: 61.8 -> 67.3 | Improved |
| evolved vs seed data at equal sample count (236 vs 236) (ablation) | SSLogic-Evolve · logic and math averages | average delta: +0.8 logic, +1.2 math | Improved |
| add SSLogic-Evolve data to math (AIME) training (ablation) | AIME-trained Qwen3-8B-Base · AIME25 | pass@1: 15.2 -> 22.4 (step 240) | Improved |
| add ARC-AGI data to the math + SSLogic-Evolve mix (ablation) | AIME + Evolve training · math benchmarks | pass@1 average delta: -3.9 to -5.0 at steps 160-200 (-3.7 at 240) | Hurt |
| train Qwen3-8B (Thinking) on Evolve instead of Seed (ablation) | Qwen3-8B (Thinking) · SynLogic | pass@1: 49.4 -> 52.1 | Improved |
| train Qwen3-8B (Thinking) on Evolve instead of Seed (ablation) | Qwen3-8B (Thinking) · math benchmarks (AIME24, Brumo25, HMMT25) | pass@1: AIME24 74.3 -> 74.1; Brumo25 71.4 -> 70.9 | No clear effect |
| evolve data, token-matched comparison (about 1.09B tokens) (ablation) | SSLogic-Evolve vs Seed · Macro Accuracy | relative macro accuracy: +7.3% relative | Improved |
| evolved data effect on KORBench categories (ablation) | Evolve vs Seed · KORBench | pass@1: Logic 32.4 -> 45.6; Puzzle 7.2 -> 8.4 | Mixed |
| mixing math with Evolve data raises reflection token rate | RL training dynamics · reflection tokens per 100 response tokens | frequency: 17.57 -> 27.80 | Improved |
| evolved vs seed data on Qwen3-8B-Base ARC-AGI (ablation) | Evolve vs Seed · ARC-AGI | pass@1: 3.1 -> 3.1 (step 240) | No clear effect |
Closest papers by meaning
- ReasoningBank: Scaling Agent Self-Evolving with Reasoning Memory (ICLR 2026 · similarity 0.90)
- T1: Advancing Language Model Reasoning through Reinforcement Learning and Inference Scaling (ICML 2025 · similarity 0.89)
- AutoTool: Dynamic Tool Selection and Integration for Agentic Reasoning (ICML 2026 · similarity 0.89)
- Deep Reasoning in General Purpose Agents via Structured Meta-Cognition (NeurIPS 2026 · similarity 0.89)
- SwS: Self-aware Weakness-driven Problem Synthesis in Reinforcement Learning for LLM Reasoning (NeurIPS 2025 · similarity 0.89)
- ReVeal: Self-Evolving Code Agents via Reliable Self-Verification (ICLR 2026 · similarity 0.89)
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.
