[2602.02742] Entropy-Guided Dynamic Tokens for Graph-LLM Alignment in Molecular Understanding

Computer Science > Machine Learning arXiv:2602.02742 (cs) [Submitted on 2 Feb 2026 (v1), last revised 2 Mar 2026 (this version, v3)] Title:Entropy-Guided Dynamic Tokens for Graph-LLM Alignment in Molecular Understanding Authors:Zihao Jing, Qiuhao Zeng, Ruiyi Fang, Yan Sun, Boyu Wang, Pingzhao Hu View a PDF of the paper titled Entropy-Guided Dynamic Tokens for Graph-LLM Alignment in Molecular Understanding, by Zihao Jing and 5 other authors View PDF Abstract:Molecular understanding is central to advancing areas such as scientific discovery, yet Large Language Models (LLMs) struggle to understand molecular graphs effectively. Existing graph-LLM bridges often adapt the Q-Former-style connector with fixed-length static tokens, which is originally designed for vision tasks. These designs overlook stereochemistry and substructural context and typically require costly LLM-backbone fine-tuning, limiting efficiency and generalization. We introduce EDT-Former, an Entropy-guided Dynamic Token Transformer that generates tokens aligned with informative molecular patches, thereby preserving both local and global structural features for molecular graph understanding. Beyond prior approaches, EDT-Former enables alignment between frozen graph encoders and LLMs without tuning the LLM backbone (excluding the embedding layer), resulting in computationally efficient finetuning, and achieves stateof-the-art results on MoleculeQA, Molecule-oriented Mol-Instructions, and property prediction bench...