Atom-anchored chain-of-thought prompting enables general LLMs to reach ≥90% accuracy on reaction-site identification and ≥74% on reactant prediction for single-step retrosynthesis on benchmarks and drug-like molecules.
42* **Process:** Expand the input into a flat list by creating a **new, separate entry for each reaction** associated with a transformation
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Atom-anchored LLMs speak Chemistry: A Retrosynthesis Demonstration
Atom-anchored chain-of-thought prompting enables general LLMs to reach ≥90% accuracy on reaction-site identification and ≥74% on reactant prediction for single-step retrosynthesis on benchmarks and drug-like molecules.