Pre-trained knowledge elevates large language models beyond traditional chemical reaction optimizers

TL;DR

Pre-trained large language models significantly improve chemical reaction optimization by enabling more effective exploration of complex, high-dimensional parameter spaces, surpassing traditional Bayesian methods especially in solution-scarce scenarios.

Contribution

This work demonstrates that pre-trained knowledge in LLMs enhances optimization in chemistry, providing a new paradigm that outperforms Bayesian optimization in complex categorical spaces.

Findings

LLMs match or outperform Bayesian optimization in most datasets.

LLMs maintain higher exploration entropy than Bayesian methods.

Pre-trained domain knowledge enables better navigation of chemical parameter space.

Abstract

Modern optimization in experimental chemistry employs algorithmic search through black-box parameter spaces. Here we demonstrate that pre-trained knowledge in large language models (LLMs) fundamentally changes this paradigm. Using six fully enumerated categorical reaction datasets (768-5,684 experiments), we benchmark LLM-guided optimization (LLM-GO) against Bayesian optimization (BO) and random sampling. Frontier LLMs consistently match or exceed BO performance across five single-objective datasets, with advantages growing as parameter complexity increases and high-performing conditions become scarce (<5% of space). BO retains superiority only for explicit multi-objective trade-offs. To understand these contrasting behaviors, we introduce a topology-agnostic information theory framework quantifying sampling diversity throughout optimization campaigns. This analysis reveals that LLMs maintain systematically higher exploration Shannon entropy than BO across all datasets while achieving superior performance, with advantages most pronounced in solution-scarce parameter spaces where high-entropy exploration typically fails-suggesting that pre-trained domain knowledge enables more effective navigation of chemical parameter space rather than replacing structured exploration strategies. To enable transparent benchmarking and community validation, we release Iron Mind (https://gomes.andrew.cmu.edu/iron-mind), a no-code platform for side-by-side evaluation of human, algorithmic, and LLM optimization campaigns with public leaderboards and complete trajectories. Our findings establish that LLM-GO excels precisely where traditional methods struggle: complex categorical spaces requiring domain understanding rather than mathematical optimization.