VibeProteinBench is a new benchmark evaluating LLMs on open-ended language-interfaced protein design across recognition, engineering, and generation, with no model showing strong performance in all areas.
Liveproteinbench: A contamination-free benchmark for assessing models’ specialized capabilities in protein science
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
citation-role summary
background 2
citation-polarity summary
years
2026 2verdicts
UNVERDICTED 2roles
background 1polarities
background 1representative citing papers
Averaging top-K attention heads selected from 10 proteins recovers contacts in one forward pass and matches or exceeds the Categorical Jacobian on bidirectional protein LMs.
citing papers explorer
-
VibeProteinBench: An Evaluation Benchmark for Language-interfaced Vibe Protein Design
VibeProteinBench is a new benchmark evaluating LLMs on open-ended language-interfaced protein design across recognition, engineering, and generation, with no model showing strong performance in all areas.
-
Protein contacts are already in the attention: a single-forward-pass alternative to the Categorical Jacobian
Averaging top-K attention heads selected from 10 proteins recovers contacts in one forward pass and matches or exceeds the Categorical Jacobian on bidirectional protein LMs.