AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Task-Awareness Improves LLM Generations and Uncertainty

Tim Tomov, Dominik Fuchsgruber, Stephan Günnemann

ICML 2026 · Large language models › Language, safety and interpretability · approach: Statistical frameworks for LLM evaluation · 0 citations (OpenAlex)

Read the paper

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
decode with latent-space Bayes-optimal response instead of beam searchLLM decoding · MAQA sets (block 1)Hamming distance: 0.892 -> 0.729Improved
decode with Bayes-optimal response instead of lowest-risk sample l_SampleLLM decoding · MAQA sets (block 1)Hamming distance: 0.838 -> 0.729Improved
decode with Bayes-optimal response instead of majority-vote l_MAPLLM decoding · MAQA sets (block 3)Hamming distance: 0.913 -> 0.858Improved
decode with Bayes response instead of beam searchLLM decoding · CNN/DailyMail summarization (block 1)structural Hamming: 2.176 -> 1.603Improved
decode with Bayes response instead of beam searchLLM decoding · MAQA probability simplex (block 2)KL divergence: 7.792 -> 5.632Improved
decode with Bayes response instead of beam searchLLM decoding · WMT19 FI-EN (block 1)cosine distance: 0.131 -> 0.124Improved
decode with Bayes response instead of beam searchLLM decoding · TriviaQA classification (block 1)exact match: 0.627 -> 0.632No clear effect
increase number of MC samples for Bayes-optimal decodingl_Bayes · MAQA setsHamming distanceImproved
use Bayes risk as uncertainty instead of semantic entropy and other estimatorsBayes risk uncertainty · 4 of 5 latent structuresPRRImproved
use Bayes risk uncertainty on single-label classificationBayes risk uncertainty · TriviaQA classificationPRRNo clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy