AI Paper Atlas

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

When Single Answer Is Not Enough: Rethinking Single-Step Retrosynthesis Benchmarks for LLMs

Bogdan Zagribelnyy, Ivan Ilin, Maksim Kuznetsov, Nikita Bondarev, Roman Schutski, Thomas MacDougall, Rim Shayakhmetov, Zulfat Miftahutdinov, Mikolaj Mizera, Vladimir Aladinskiy, Alex Aliper, Alex Zhavoronkov

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
train on CREED instead of USPTO-full (ablation)C3LM-LFM2 (LFM2 2.6B SFT) · URSA-expert-2026Av. PT-Top-10 CC: 0.31 -> 0.38Improved
train on CREED instead of USPTO-full (ablation)C3LM-LFM2 (LFM2 2.6B SFT) · USPTO-50K-testAv. PT-Top-10 CC: 0.57 -> 0.45Hurt
add USPTO-full to CREED training (ablation)C3LM-LFM2-CREED · URSA-expert-2026Av. PT-Top-10 CC: 0.38 -> 0.40Improved
add USPTO-full to CREED training (ablation)C3LM-LFM2-CREED · URSA-expert-2026unique valid reactant sets: 58% -> 42%Hurt
add USPTO-full to CREED training (ablation)C3LM-LFM2-CREED · USPTO-50K-testAv. PT-Top-10 CC: 0.45 -> 0.65Improved
add GRPO RFT with Molecular Transformer reward (ablation)C3LM-LFM2-CREED+USPTO · URSA-expert-2026Av. PT-Top-10 CC: 0.40 -> 0.43Improved
add GRPO RFT with CREED multiple-answer exact-match reward (ablation)C3LM-LFM2-CREED+USPTO · URSA-expert-2026Av. PT-Top-10 CC: 0.40 -> 0.40No clear effect
add GRPO RFT with Molecular Transformer reward (ablation)C3LM-LFM2-CREED+USPTO · USPTO-50K-testAv. PT-Top-10 CC: 0.65 -> 0.68Improved
evaluate on novel URSA-expert-2026 instead of USPTO-50K-testLLMs overall · URSA vs USPTO-50KAv. PT-Max CC: Claude 4.5 Sonnet 3.37 -> 1.45Hurt
fine-tune small LFM2 2.6B on CREED+USPTOLFM2 2.6B · URSA-expert-2026Av. PT-Top-10 CC: 0.00 -> 0.40Improved
use open-weight general LLMsQwen3 8B / DeepSeek 3.2 · URSA-expert-2026Av. PT-Top-10 CC: 0.00 and 0.05Hurt
use LLMs rather than conventional SSRS modelsLLMs · URSA-expert-2026ChemCensor plausibilityHurt

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