AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

MultiHal: Multilingual Dataset for Knowledge-Graph Grounded Evaluation of LLM Hallucinations

Ernests Lavrinovics, Russa Biswas, Katja Hose, Johannes Bjerva

ICML 2026 · Large language models › Retrieval and knowledge · approach: Knowledge-augmented question answering · 1 citations (OpenAlex)

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
add KG paths as in-context knowledge (KG-RAG) instead of vanilla QAGPT-4o Mini · MultiHal NLI, all languagesentailment (%): 40% -> 76%Improved
add KG paths as in-context knowledgeGemini 2.0 Flash · MultiHal NLI, all languagesentailment (%): 52% -> 68%Improved
add KG paths as in-context knowledgeLlama 3.3 70B Instruct · MultiHal NLI, all languagesentailment (%): 45% -> 71%Improved
add KG paths as in-context knowledgeGPT-4o Mini · MultiHal English, HHEM-2.1hallucinated (%): 53% -> 11%Improved
add KG paths as in-context knowledgeGemini 2.0 Flash · MultiHal English, HHEM-2.1hallucinated (%): 42% -> 13%Improved
add KG paths as in-context knowledgeLlama 3.3 70B Instruct · MultiHal English, HHEM-2.1hallucinated (%): 45% -> 12%Improved
restrict KG paths to quality rating 4 only versus vanilla QA (ablation)GPT-4o Mini · MultiHal semantic similaritysemantic similarity: 0.4 -> 0.52Improved
use GPT-4o Mini instead of Gemini 2.0 Flash as KG-path judge (ablation)KG path filtering · English subset, KG-RAGcorrelation of path score with semantic score: 0.430 -> 0.485Improved
use GPT-4o Mini instead of Gemini 2.0 Flash as KG-path judge (ablation)KG path filtering · English subset, KG-RAGsemantic score: 0.529 -> 0.513No clear effect
separate KG path labels with semicolons before translation (ablation)NLLB-200 translation · translation of KG pathstranslation qualityImproved
add KG paths for suggestive, temporal or reasoning-heavy subsetsKG-RAG · TruthfulQA, HaluBench, Felm subsetssemantic similarityMixed

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