Searcharxiv⌕ Search

arXiv · 2611.22623

What Do LLMs Use to Judge Synthesis Recipes? An Experimental Analysis on 240 Solid-State Syntheses

Abstract

Inorganic materials discovery is increasingly limited by synthesis rather than by material design. Large language models (LLMs) are being adopted to propose synthesis conditions with little experimental evidence of their capabilities and biases. Here, we carry out an experimental synthesis benchmark of 240 experiments across eight known and unknown inorganic oxides and subsequently evaluate an LLM as a zero-shot ranker over fixed banks of solid-state synthesis recipes. For each pair of recipes for the same oxide target, `GPT-5.6-Sol` selects the recipe expected to yield higher target-phase purity. We aggregate these pairwise preferences using a regularized Bradley-Terry model to obtain a ranked list of synthesis recipes for a given target material. By comparison with the actual purity, obtained through X-ray diffraction, we measure `GPT-5.6-Sol`'s performance as a zero-shot ranker (0.72 pairwise accuracy). By selective removal of the information exposed to the model during this ranking, we conclude that ranking is driven predominantly by thermal features, with reaction chemistry modulating the effect. Ranking accuracy tracks the purity-temperature correlation by target (Spearman ρ= 0.91\), \(p = 0.002\), and on the six thermally-driven targets withholding all chemical identity hardly impacts performance (0.68) compared to withholding temperature (0.56). Ranking performance was higher on known than unknown targets, but the group-level difference was not statistically conclusive. The methodological choices highlight the need for more experimentally validated synthesis benchmarks and position LLMs as a thermally informed prior.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Derek van Tilborg, Laura-Bianca Pasca, Harry Sevi, Kyle Reeves, Thomas Loux, Aikaterini Flessa Savvidou, Varincka Jean-Gilles, Marilou Carouge, Joonatan E. M. Laulainen. 2026-11-17. What Do LLMs Use to Judge Synthesis Recipes? An Experimental Analysis on 240 Solid-State Syntheses. https://searcharxiv.com/abs/2611.22623

Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite Searcharxiv by Indiciae: BibTeX · RIS.

KEEP EXPLORING

Related papers

Adaptive Finite State Projection with Quantile-Based Pruning for Solving the Chemical Master Equation

We present an adaptive Finite State Projection (FSP) method for efficiently solving the Chemical Master Equation (CME) with rigorous error control. Our approach integrates time-stepping with dynamic state-space truncation, balancing accuracy and computational cost. Krylov subspace methods approximate the matrix exponential, while quantile-based pruning controls state-space growth by removing low-probability states. Theoretical error bounds ensure that the truncation error remains bounded by the pruned mass at each step, which is user-controlled, and does not propagate forward in time. Numerical experiments on biochemical systems, including the Lotka-Volterra and Michaelis-Menten and bi-stable toggle switch models.

cs.CE↗

SoK: Market Microstructure for Decentralized Prediction Markets (DePMs)

Decentralized prediction markets (DePMs) allow open participation in event-based wagering without fully relying on centralized intermediaries. We review the history of DePMs which date back to 2011 and includes hundreds of proposals. Perhaps surprising, modern DePMs like Polymarket deviate materially from earlier designs like Truthcoin and Augur v1. We use our review to present a modular workflow comprising eight stages: underlying infrastructure, market topic, share structure and pricing, market initialization, trading, market resolution, settlement, and archiving. For each module, we enumerate the design variants, analyzing trade-offs around decentralization, expressiveness, and manipulation resistance. We also identify open problems for researchers interested in this ecosystem.

cs.CE↗

A Multigrid Finite State Projection Algorithm For The Solution Of The Reaction-Diffusion Master Equation

We develop a multigrid method for the finite state projection of the reaction-diffusion master equation (RDME). The RDME evolves the joint probabilities of molecular count configurations across well-mixed voxels coupled by local reactions and diffusive jumps. When fast diffusion or large reaction rates make this evolution stiff over long time horizons, implicit time stepping requires solving a large shifted linear system at every step. Grouping neighboring voxels and summing their molecular counts produces a hierarchy of coarser state spaces terminating at a one-voxel chemical master equation. Restriction sums residual entries with the same coarse counts, while multinomial prolongation preserves the RDME generator form on every level with explicit coarse reaction and diffusion rates. For reversible diffusion, the error omitted by coarsening is bounded by the graph Dirichlet energy of jumps inside the aggregates, with a constant independent of population. Numerical tests with standard multigrid cycles and point or block Gauss-Seidel smoothing reach 7.16 million states and resolve rare probabilities of order $10^{-7}$.

cs.CE↗