# ChemATP: A Training-Free Chemical Reasoning Framework for Large Language Models

arXiv: https://arxiv.org/abs/2512.19240 · HTML: https://mlsanity.com/paper/2512.19240/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Clinical and biomedical applications, one of 17 directions in LLM reasoning.
- Builds on 20 of its 47 references that are in the map.
- First posted on arXiv 22 Dec 2025.

## Placed in

- LLM reasoning / [Clinical and biomedical applications](https://mlsanity.com/lr/clinical-and-biomedical-applications/)

## Builds on (most cited first, 20 of 20)

- [Semi-Supervised Classification with Graph Convolutional Networks](https://mlsanity.com/paper/1609.02907/) (2016)
- [Graph Attention Networks](https://mlsanity.com/paper/1710.10903/) (2017)
- [How Powerful are Graph Neural Networks?](https://mlsanity.com/paper/1810.00826/) (2018)
- [Deep Graph Infomax](https://mlsanity.com/paper/1809.10341/) (2018)
- [Strategies for Pre-training Graph Neural Networks](https://mlsanity.com/paper/1905.12265/) (2019)
- [Directional Message Passing for Molecular Graphs](https://mlsanity.com/paper/2003.03123/) (2020)
- [InfoGraph: Unsupervised and Semi-supervised Graph-Level Representation Learning via Mutual Information Maximization](https://mlsanity.com/paper/1908.01000/) (2019)
- [Draft, Sketch, and Prove: Guiding Formal Theorem Provers with Informal Proofs](https://mlsanity.com/paper/2210.12283/) (2022)
- [A Molecular Multimodal Foundation Model Associating Molecule Graphs with Natural Language](https://mlsanity.com/paper/2209.05481/) (2022)
- [MolCA: Molecular Graph-Language Modeling with Cross-Modal Projector and Uni-Modal Adapter](https://mlsanity.com/paper/2310.12798/) (2023)
- [LEGO-Prover: Neural Theorem Proving with Growing Libraries](https://mlsanity.com/paper/2310.00656/) (2023)
- [InstructMol: Multi-Modal Integration for Building a Versatile and Reliable Molecular Assistant in Drug Discovery](https://mlsanity.com/paper/2311.16208/) (2023)
- [MolXPT: Wrapping Molecules with Text for Generative Pre-training](https://mlsanity.com/paper/2305.10688/) (2023)
- [Developing ChemDFM as a large language foundation model for chemistry](https://mlsanity.com/paper/2401.14818/) (2024)
- [UniMoT: Unified Molecule-Text Language Model with Discrete Token Representation](https://mlsanity.com/paper/2408.00863/) (2024)
- [Motif-aware Attribute Masking for Molecular Graph Pre-training](https://mlsanity.com/paper/2309.04589/) (2023)
- [Uni-Mol Docking V2: Towards Realistic and Accurate Binding Pose Prediction](https://mlsanity.com/paper/2405.11769/) (2024)
- [Can LLMs Reason Over Non-Text Modalities in a Training-Free Manner? A Case Study with In-Context Representation Learning](https://mlsanity.com/paper/2509.17552/) (2025)
- [AtomDisc: An Atom-level Tokenizer that Boosts Molecular LLMs and Reveals Structure--Property Associations](https://mlsanity.com/paper/2512.03080/) (2025)
- [FragmentNet: Adaptive Graph Fragmentation for Graph-to-Sequence Molecular Representation Learning](https://mlsanity.com/paper/2502.01184/) (2025)

## Abstract (arXiv)

Large Language Models (LLMs) exhibit strong general reasoning but struggle in molecular science due to the lack of explicit chemical priors in standard string representations. Current solutions face a fundamental dilemma. Training-based methods inject priors into parameters, but this static coupling hinders rapid knowledge updates and often compromises the model's general reasoning capabilities. Conversely, existing training-free methods avoid these issues but rely on surface-level prompting, failing to provide the fine-grained atom-level priors essential for precise chemical reasoning. To address this issue, we introduce ChemATP, a framework that decouples chemical knowledge from the reasoning engine. By constructing the first atom-level textual knowledge base, ChemATP enables frozen LLMs to explicitly retrieve and reason over this information dynamically. This architecture ensures interpretability and adaptability while preserving the LLM's intrinsic general intelligence. Experiments show that ChemATP significantly outperforms training-free baselines and rivals state-of-the-art training-based models, demonstrating that explicit prior injection is a competitive alternative to implicit parameter updates.