AI Paper Atlas

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

FlowBot: Inducing LLM Workflows with Bilevel Optimization and Textual Gradients

Hongyeon Yu, Young-Bum Kim, Yoon Kim

ICML 2026 · Large language models › Agents, code and math · approach: Agent instruction-following benchmarks · 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
add bilevel workflow induction on top of prompt optimization (ablation)FlowBot · HotpotQA1score: 60.07 -> 72.80Improved
add bilevel optimization formulation (ablation)FlowBot · HotpotQA1score: 61.73 -> 72.80Improved
add bilevel optimization formulation (ablation)FlowBot · HotpotQA2score: 79.26 -> 80.19Improved
add layer-by-layer textual gradient backpropagation (ablation)FlowBot · HotpotQA1score: 71.13 -> 72.80Improved
add layer-by-layer textual gradient backpropagation (ablation)FlowBot · HotpotQA2score: 76.51 -> 80.19Improved
add workflow induction vs prompt-only optimization (Qwen3.5-4B executor) (ablation)Qwen3.5-4B · HotPotQA1score: 0.513 -> 0.681Improved
add workflow induction vs prompt-only optimization (Qwen3.5-397B executor) (ablation)Qwen3.5-397B-A17B · HotPotQA2score: 0.816 -> 0.806Hurt
use FlowBot instead of GEPAGPT-4.1 mini · HoVerscore: 51.67 -> 63.20Improved
use FlowBot instead of GEPA+MergeGPT-4.1 mini · PUPAscore: 96.46 -> 90.67Hurt
use FlowBot instead of reimplemented AFlowGPT-4.1 mini · DROPscore: 79.69 -> 92.28Improved
use FlowBot instead of reimplemented AFlowGPT-4.1 mini · MBPPscore: 74.78 -> 73.84Hurt
induce one unified workflow instead of separate onesFlowBot Qwen3.5-397B · 6 AFlow tasks averagescore: 87.85 -> 80.48Hurt

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