Code and lightweight audit artifacts for auditable repair of scientific reasoning graph extraction on a 350-row benchmark.
-
Updated
Jul 21, 2026 - Python
Code and lightweight audit artifacts for auditable repair of scientific reasoning graph extraction on a 350-row benchmark.
Replication package for “Illocutionary Explanation Planning for Source-Faithful Explanations in Retrieval-Augmented Language Models” (xAI 2026): code, data, prompts, and evaluation for chain-of-illocution prompting in textbook-grounded RAG.
Audit whether an LLM's output is actually grounded in its source — atomic-claim verification, NLI cross-check, and agreement scoring against human labels
Linear probes on Llama 3.1 8B hidden states rank claim correctness better than the model's stated confidence; legal QA and biography probes, Modal runs. Write-up: casella.dev/blog_interp.html
To associate your repository with the factscore topic, visit your repo's landing page and select "manage topics."