跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.20848 · v1

MolReps: Taking SELFIES, Making SMILES, Teaching LLMs

首次提交 2026-11-16 18:23
版本更新 2026-11-16 18:23
雷达观测 2026-11-17 17:11
arXiv ↗PDF ↗

摘要

The same molecule can be expressed in multiple, syntactically distinct representations, yet it remains unclear whether large language models (LLMs) can recognize and relate these equivalent encodings. We introduce MolReps, a pipeline that generates 9,600 question-answer pairs across ten representations for conversion, normalization, and 2D structure recognition. Post-training on MolReps improves text-only and multimodal models, MolQwen3-4B-Instruct (MQ3-4B-I) and MolQwen3-VL-4B-Instruct (MQ3-VL-4B-I), from 4.01%4.01\% to 75.13%75.13\% and from 7.11%7.11\% to 73.14%73.14\% (InChIKey match) on the molecule-overlap test set, respectively. On the molecule-disjoint test set, post-training increased validity from 20.35%20.35\% to 58.12%58.12\% for MQ3-4B-I and from 17.19%17.19\% to 59.34%59.34\% for MQ3-VL-4B-I. However, InChIKey match remained low, decreasing from 3.86%3.86\% to 3.09%3.09\% and from 6.44%6.44\% to 2.29%2.29\%, respectively. Furthermore, the resulting models achieve a relative 27.84%27.84\% improvement in average over their base model on MolRepBench, a comprehensive benchmark for evaluating molecular property prediction. These results show the effectiveness of targeted post-training for enhancing molecular representation understanding and its potential to improve downstream molecular tasks.

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-16 18:23 · 最新观测版本

MolReps: Taking SELFIES, Making SMILES, Teaching LLMs

The same molecule can be expressed in multiple, syntactically distinct representations, yet it remains unclear whether large language models (LLMs) can recognize and relate these equivalent encodings. We introduce MolReps, a pipeline that generates 9,600 question-answer pairs across ten representations for conversion, normalization, and 2D structure recognition. Post-training on MolReps improves text-only and multimodal models, MolQwen3-4B-Instruct (MQ3-4B-I) and MolQwen3-VL-4B-Instruct (MQ3-VL-4B-I), from 4.01%4.01\% to 75.13%75.13\% and from 7.11%7.11\% to 73.14%73.14\% (InChIKey match) on the molecule-overlap test set, respectively. On the molecule-disjoint test set, post-training increased validity from 20.35%20.35\% to 58.12%58.12\% for MQ3-4B-I and from 17.19%17.19\% to 59.34%59.34\% for MQ3-VL-4B-I. However, InChIKey match remained low, decreasing from 3.86%3.86\% to 3.09%3.09\% and from 6.44%6.44\% to 2.29%2.29\%, respectively. Furthermore, the resulting models achieve a relative 27.84%27.84\% improvement in average over their base model on MolRepBench, a comprehensive benchmark for evaluating molecular property prediction. These results show the effectiveness of targeted post-training for enhancing molecular representation understanding and its potential to improve downstream molecular tasks.

查看此版本解读 →查看此版本原文 ↗