AI Paper Atlas

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

Agentic Context Engineering: Learning Comprehensive Contexts for Self-Improving Language Models

Qizheng Zhang, Changran Hu, Shubhangi Upasani, Boyuan Ma, Fenglu Hong, Vamsidhar Kamanuru, Jay Rainton, Chen Wu, Mengmeng Ji, Hanchen Li, Urmish Thakker, James Zou and 1 more

ICLR 2026 · Large language models › Agents, code and math · approach: Agent instruction-following benchmarks

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 incremental delta context updates (ablation)ACE · AppWorld test-normalaverage score: 56.9 -> 70.3Improved
add Reflector iterative refinement (ablation)ACE · AppWorldaverage score: 55.1 -> 56.8Improved
add multi-epoch adaptation (ablation)ACE · AppWorldaverage score: 56.8 -> 59.4Improved
add offline warmup before online adaptation (ablation)ACE · AppWorld onlineaverage score: 56.1 -> 59.5Improved
replace GEPA with ACE for offline adaptationReAct · AppWorldaverage score: 46.4 -> 59.4Improved
remove ground-truth labels from the ReflectorACE offline · AppWorldaverage score: 59.4 -> 57.2No clear effect
remove ground-truth labels from the ReflectorACE online · Financial benchmarkaverage accuracy: 76.6 -> 72.9Hurt
remove ground-truth labels from the ReflectorDC (CU) online · Financial benchmarkaverage accuracy: 71.8 -> 65.4Hurt
replace Dynamic Cheatsheet with ACE onlineReAct · AppWorldaverage score: 51.9 -> 59.5Improved
replace GEPA with ACE for adaptation costAppWorld offline · adaptation latencyseconds: 53898 -> 9517Improved
inject a harmful Reflector update every iteration (ablation)ACE · FiNERaccuracy: 78.3 -> 66.7Hurt
increase reflection iterations from 5 to 10 (ablation)ACE · AppWorld test-normalaverage score: 67.6 -> 65.2Hurt

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