AI Paper Atlas

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

A Systematic Study of Behavioral Cloning for Scientific Data Annotation

Ishaan Singh Chandok, Core Francisco Park

ICML 2026 · Large language models › Agents, code and math · approach: LLMs for scientific discovery · 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
Pretrain on 9 tasks then fine-tune, versus training from scratch (ablation)Multi-task BC model · SHAPE MATCHING (held-out)accuracy: 0% -> 76.6%Improved
Provide in-context demonstrations or action prefixes instead of fine-tuning (ablation)Pretrained multi-task model · SHAPE MATCHING (held-out)accuracy: 0%No clear effect
Increase fine-tuning data from 500 to 7800 sequences (ablation)Pretrained multi-task model · SHAPE MATCHINGaccuracy: 76.6% at 500; similar at 7800No clear effect
Scale model parameters 10x (25M to 320M) (ablation)Multi-task BC · Training loss vs tokenssample efficiency: roughly 3x better sample efficiencyImproved
Scale model from 25M to 320M at equal FLOPs (ablation)Multi-task BC · Training loss at equal FLOPslossHurt
Larger model at equal training loss (ablation)Multi-task BC · Decision actions (done, undo, navigate)accuracyHurt
Larger model at equal training loss (ablation)Multi-task BC · Motor actionsaccuracyNo clear effect
Use the 95M BC model instead of prompting VLMsBC vs Gemini 3 Flash / Qwen3-VL-32B · COLORED DOT TRACKING placement @5pxaccuracy: 97.4% vs 80.0% / 25.0%Improved
Apply beta-DAgger on failing tasks (ablation)BC model · Animal behavioral and limb trackingautoregressive successNo clear effect
Fine-tune briefly on 8 GUI layout variantsPretrained BC model · COLORED DOT TRACKING GUI variantsperformanceImproved
Train on data containing mistakes then undoBC model · Generative mistake ratemistake rate: 9.2% in data vs 1.0% generatedMixed

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