Beyond Benchmark Accuracy: A Diagnostic Test for LLM Chemical Prediction

arXiv:2610.19906 · cs.CL · Submitted 2026-10-14 · Read on arXiv

cs.CL

Submitted: 2026-10-14

Updated: 2026-10-14

Comments: Under review at ICLR 2027

Code: https://github.com/pyrfume/pyrfume-data

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Accuracy — 1× in this paper · explained in 1 episode
• Benchmark — 1× in this paper · explained in 2 episodes

Sources

• MolecularIQ: Characterizing Chemical Reasoning Capabilities Through Symbolic Verification on Molecular Graphs
• Molecular D'ej`a Vu: Digit-Level Retrieval of Molecular Properties in Frontier Language Models
• In-Context Molecular Property Prediction with LLMs: A Blinding Study on Memorization and Knowledge Conflicts
• From Answers to States: Verifiable Process-Level Evaluation of Chemical Reasoning in Large Language Models
• Clever Materials: When Models Identify Good Materials for the Wrong Reasons
• Benchmark for Assessing Olfactory Perception of Large Language Models

Related papers

• Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
• Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
• Subliminal Steering: Stronger Encoding of Hidden Signals
• MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
• The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
• Untangling the Mechanisms of Misleading Context in Medical Question Answering