Deterministic, no-LLM gate for LLM output - traces every claim to an evidence bank, scores quality with a fixed rubric, and flags repetition runaway, so the check is auditable and repeatable.
-
Updated
Aug 29, 2026 - Python
#
ragtruth
Here are 5 public repositories matching this topic...
Empirical evaluation and error analysis of LettuceDetect for RAG hallucination detection using RAGTruth.
python nlp machine-learning ai deep-learning evaluation error-analysis rag llm hallucination-detection ragtruth
-
Updated
Aug 14, 2026 - Jupyter Notebook
A self-correcting RAG layer that cut hallucination from 35% to 13% on RAGTruth by verifying every claim against the retrieved context.
fact-checking nli rag self-correction llm faithfulness retrieval-augmented-generation llm-evaluation hallucination-detection ragtruth
-
Updated
Jun 21, 2026 - Python
A diverse 6-LLM judge committee (4 labs, 3 countries) has correlated errors: ~1.9 effective independent judges of 6 (mean pairwise error-correlation ρ≈0.42) on RAGTruth. Provider diversity does not buy error-independence. Verdicts + one-command recompute included.
consensus ai-safety model-ensemble llm llm-evaluation hallucination-detection llm-as-judge correlated-errors ragtruth
-
Updated
Aug 19, 2026 - Python
Training-free committee-of-6 LLM judge for RAG hallucination detection: AUROC 0.88 / F1 0.77 on RAGTruth, reproducible at $0 from frozen verdicts — and it surfaces provable label errors in the benchmark.
-
Updated
Aug 2, 2026 - Python
Add this topic to your repo
To associate your repository with the ragtruth topic, visit your repo's landing page and select "manage topics."