MolPuzzle Benchmark

What is MolPuzzle?

MolPuzzle is a benchmark for evaluating MLM's reasoning ability, comprising 234 instances of structure elucidation, which feature over 18,000 QA samples.

Released

2024-09-26

Evaluates

推理, Reasoning, NeurIPS 2024, 科学智能, AI for Science, 逻辑推理, 科学推理, Scientific Reasoning, 不支持, Unsupported

Openness

open

Importer

OpenCompass Hub

Reported scores

0

Source provenance

• Original evidence https://hub.opencompass.org.cn/dataset-detail/MolPuzzle

MolPuzzle code and dataset

• Code https://github.com/KehanGuo2/MolPuzzle
• Dataset https://huggingface.co/datasets/kguo2/MolPuzzle_data

Which sources cite MolPuzzle?

• MolPuzzle

Related 推理, Reasoning, NeurIPS 2024, 科学智能, AI for Science, 逻辑推理, 科学推理, Scientific Reasoning, 不支持, Unsupported benchmarks

• A-Bench
• A-OKVQA
• AbsPyramid
• ActionAtlas
• ActiveFly-Bench
• Ada-LEval
• AdvancedMathBench
• AECBench