AI Paper Atlas

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

Meta Context Engineering via Agentic Skill Evolution

Haoran Ye, Xuning He, Vincent Arak, Haonan Dong, Guojie Song

ICML 2026 · Large language models › Agents, code and math · approach: Long-horizon LLM agents · 1 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
evolve skills across iterations instead of skill-less base agent (ablation)MCE · FiNER (offline)accuracy (%): 73.0 -> 75.0Improved
evolve skills versus a single fixed skill from the task spec (ablation)MCE · FiNER (offline)accuracy (%): 71.0 -> 75.0Improved
add a fixed skill to the base agent (online) (ablation)MCE · FiNER (online)accuracy (%): 67.0 -> 68.0No clear effect
add a fixed skill to the base agent (offline) (ablation)MCE · FiNER (offline)accuracy (%): 73.0 -> 71.0Hurt
fully agentic base agent without any manual scaffolding versus ACE (ablation)MCE (w/o skills) · FiNERaccuracy (%): 71.0 -> 73.0 offline; 64.0 -> 67.0 onlineImproved
use MiniMax M2.1 as ACE reflector/curator instead of DeepSeek V3.1 (ablation)ACE · FiNERfinal accuracy (%): 70 -> 67Hurt
replace ACE's additive curation with MCE's agentic batch-level optimizationcontext learning · FiNERtraining time (hours): 25.8h -> 1.9h (13.6x faster)Improved
use MCE instead of ACE for rollout efficiencycontext learning · FiNERrollouts to reach training accuracy: 450 rollouts for 95% vs ACE peak 94% at 2169Improved
use MCE contexts instead of ACE contexts for transfer to Llama3.3-70Btransferred context · LawBench (transfer to Llama3.3-70B)micro-F1: ACE 0.19 (base 0.24) vs MCE 0.27Improved
use MCE instead of ACE for the Symptom2Disease online settingcontext learning · Symptom2Disease (online)accuracy (%): 62.3 -> 76.4Improved
ACE reflection-curation on a task where raw examples sufficeACE · Symptom2Disease (offline)accuracy (%): ICL 84.4 vs ACE 79.2Hurt
MCE versus GEPA on a lightweight generatorMCE · Aegis2.0 (Qwen3-8B)F1: 0.76 -> 0.80Improved

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