AI Paper Atlas

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

Embedding Trust: Semantic Isotropy Predicts Nonfactuality in Long-Form Text Generation

Dhrupad Bhardwaj, Julia Kempe, Tim G. J. Rudner

ICML 2026 · Large language models › Language, safety and interpretability · approach: Faithfulness of LLM explanations

Read the paperCheck your idea against 137,217 papers

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
replace LUQ with semantic isotropy (Gemini)uncertainty scoring · Book Summaries, Phi 3.5 MiniR^2: 0.33 -> 0.43Improved
replace LUQ with semantic isotropy (Mistral E5)uncertainty scoring · FactScore-Bio, GPT 4.1 MiniR^2: 0.12 -> 0.47Improved
replace LUQ with semantic isotropy (Nomic v1) on a stronger generatoruncertainty scoring · TriviaQA, GPT 4.1 MiniR^2: 0.10 -> 0.14Mixed
replace LUQ with semantic isotropy (Gemini)uncertainty scoring · TriviaQA, GPT 4.1 MiniR^2: 0.10 -> 0.03Hurt
use Cohere v3.0 isotropy instead of LUQ under FactScore labelsuncertainty scoring · FactScore-Bio (FactScore), Llama 3.1 8BR^2: 0.657 -> 0.75Improved
replace von Neumann entropy with Frobenius norm (ablation)semantic isotropy score (Gemini) · TriviaQA, Phi 3.5 MiniR^2: 0.43 -> 0.41No clear effect
replace von Neumann entropy with inverse trace (ablation)semantic isotropy score (Gemini) · TriviaQA, Phi 3.5 MiniR^2: 0.38 -> 0.43Improved
replace von Neumann entropy with LogDet (ablation)semantic isotropy score (Nomic v1) · TriviaQA, Phi 3.5 MiniR^2: 0.35 -> 0.39Hurt
use large instead of small OpenAI embedding (ablation)semantic isotropy score · TriviaQA, Llama 3.1 8BR^2: 0.16 -> 0.16No clear effect
use mean pooling instead of last-token embedding for Nomic v1 (ablation)semantic isotropy score · poolingR^2Improved
increase Qwen 3 embedding size from 0.6B to 8B (ablation)semantic isotropy score · Book Summaries, Llama 3.1 8BR^2: 0.18 -> 0.33Improved
increase Qwen 3 embedding size from 0.6B to 8B (ablation)semantic isotropy score · TriviaQA, Llama 3.1 8BR^2: 0.05 -> 0.00Hurt

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