AI Paper Atlas

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

On the Generalization Gap in Self-Evolving Language Model Reasoning

Zhenting Qi, Susanna Maria Baby, Stefanie Anna Baby, Kan Yuan, Andrew Tomkins, Tu Vu, Da-Cheng Juan, Cyrus Rashtchian

ICML 2026 · Large language models › Agents, code and math · approach: LLMs for optimization and heuristics · 0 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
apply SimpleSE self-evolution (tau=0.6)gemma-3-4b-it · KKaccuracy (%): 31.0 -> 40.7Improved
add multi-turn feedback revision (RevisionSE) over SimpleSEgemma-3-12b-it · KKaccuracy (%): 47.5 -> 52.8Improved
add multi-turn feedback revision (RevisionSE) (ablation)gemma-3-1b-it · KKaccuracy (%): 23.8 -> 22.4Hurt
add iterative rounds of SimpleSE (three rounds) (ablation)gemma-3-4b-it · KKaccuracy (%): 31.0 -> 44.1Improved
add curriculum (2-3 then 4-5 people) instead of random mixing (ablation)gemma-3-4b-it · KKaccuracy (%): 41.2 -> 44.8Improved
add a final oracle verification round after SimpleSE iterations (ablation)gemma-3-4b-it · KKaccuracy (%): 43.3 -> 53.2Improved
scale model size from 1B to 4B for SimpleSE (ablation)gemma-3 SimpleSE · KKaccuracy (%): 7.8 -> 8.4Improved
raise verifier threshold tau (ablation)gemma-3-4b-it SimpleSE · KKaccuracy (%): 41.7Mixed
apply SimpleSE trained on OpenThoughts3gemma-3-4b-it · MATH500accuracy (%): 75.8 -> 77.4Improved
apply SimpleSE trained on OpenThoughts3gemma-3-4b-it · GSM8Kaccuracy (%): 89.2 -> 89.0No clear effect
apply SimpleSE trained on OpenThoughts3Qwen2.5-7B-Instruct · KKaccuracy (%): 18.1 -> 17.6Hurt
apply SimpleSE iterations (ablation)gemma-3-4b-it · KK Pass@32Pass@32 (%): 78.0 -> 78.8No clear effect

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