# Can AI Scientist Agents Learn from Lab-in-the-Loop Feedback? Evidence from Iterative Perturbation Discovery

Source: [Can AI Scientist Agents Learn from Lab-in-the-Loop Feedback? Evidence from Iterative Perturbation Discovery](<https://www.sota2.com/research/paper/69ca03620636e5bf4afc9206>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-31T05:00:04.048Z

Authors: Gilles Wainrib, Barbara Bodinier, Haitem Dakhli, Josep Monserrat, Almudena Espin Perez, Sabrina Carpentier, Roberta Codato, John Klein

Published: 2026-03-27T08:48:30.000Z

## Abstract

Recent work has questioned whether large language models (LLMs) can perform genuine in-context learning (ICL) for scientific experimental design, with prior studies suggesting that LLM-based agents exhibit no sensitivity to experimental feedback. We shed new light on this question by carrying out 800 independently replicated experiments on iterative perturbation discovery in Cell Painting high-content screening. We compare an LLM agent that iteratively updates its hypotheses using experimental feedback to a zero-shot baseline that relies solely on pretraining knowledge retrieval. Access to feedback yields a $+53.4\\%$ increase in discoveries per feature on average ($p = 0.003$). To test whether this improvement arises from genuine feedback-driven learning rather than prompt-induced recall of pretraining knowledge, we introduce a random feedback control in which hit/miss labels are permuted. Under this control, the performance gain disappears, indicating that the observed improvement depends on the structure of the feedback signal ($+13.0$ hits, $p = 0.003$). We further examine how model capability affects feedback utilization. Upgrading from Claude Sonnet 4.5 to 4.6 reduces gene hallucination rates from ${\\sim}33\\%$--$45\\%$ to ${\\sim}3$--$9\\%$, converting a non-significant ICL effect ($+0.8$, $p = 0.32$) into a large and highly significant improvement ($+11.0$, $p=0.003$) for the best ICL strategy. These results suggest that effective in-context learning from experimental feedback emerges only once models reach a sufficient capability threshold.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Cumulative target feature discovery on JUMP Cell Painting](<https://www.sota2.com/research/sota/cumulative-target-feature-discovery-on-jump-cell-painting.md>): Cumulative target feature discovery; JUMP Cell Painting; F0 Score: 33.4.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2603.26177>)

[PDF](<https://arxiv.org/pdf/2603.26177.pdf>)