AI Paper Atlas

Research map / Large language models / Agents, code and math

Theoretical Modeling of Large Language Model Self-Improvement Training Dynamics Through Solver-Verifier Gap

Yifan Sun, Yushan Liang, Zhen Zhang, Xin Liu, Jiaye Teng

ICLR 2026 · Large language models › Agents, code and math · approach: Mathematical and structural reasoning benchmarks · 6 citations (OpenAlex)

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 changedWhereMeasuredEffect
add external DeepSeek-V3 data (Early) to self-improvementPhi-4-mini self-improvement · Math (train, QE)solver accuracy (%): 45.08 -> 46.33Improved
add external DeepSeek-V3 data (Early) to self-improvementPhi-4-mini self-improvement · GSM8k (train, QE)solver accuracy (%): 88.53 -> 88.54No clear effect
add external DeepSeek-V3 data (Early) to self-improvementLlama-3.2-3B self-improvement · Math (train, QE)solver accuracy (%): 49.16 -> 53.48Improved
add external DeepSeek-V3 data (Early) to self-improvementLlama-3.2-3B self-improvement · GSM8k (train, QE)solver accuracy (%): 88.66 -> 88.26No clear effect
average cross-improvement over baselinePhi-4-mini / Llama-3.2-3B · GSM8k (train, QE)solver accuracy (%): +0.10 (Phi-4-mini), -0.77 (Llama-3.2-3B)Mixed
average cross-improvement over baselineLlama-3.2-3B self-improvement · Math (train, QE)solver accuracy (%): +3.35Improved
vary external-data timing: Early vs Late (ablation)Llama-3.2-3B cross-improvement · Math (train, QE)solver accuracy (%): 53.48 vs 51.31Mixed
vary external-data timing: Early vs Uniform vs Late (ablation)Phi-4-mini cross-improvement · Math (train, QE)solver accuracy (%): Max-Min 0.73No clear effect
vary external-data timing: Early vs Uniform vs Late (ablation)Phi-4-mini cross-improvement · GSM8k (train, QE)solver accuracy (%): Max-Min 0.46No clear effect
vary external-data timing: Early vs Uniform vs Late (ablation)Llama-3.2-3B cross-improvement · GSM8k (train, QE)solver accuracy (%): Max-Min 1.07No clear effect
early external data raises initial verifier accuracy (ablation)Phi-4-mini · Math (train, QE)solver accuracy (%): +5.87Improved
early external data raises initial verifier accuracy (ablation)Llama-3.2-3B · GSM8k (train, QE)solver accuracy (%): -1.58Hurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy