Longterm Wiki
Updated 2026-03-13HistoryData
Page StatusDocumentationDashboard
Edited today
Content0/12
LLM summaryScheduleEntityEdit history
Tables0Diagrams0Int. links0/ ~5Ext. links0Footnotes0References0Quotes0Accuracy0

Hallucination Evals

Eval results from the hallucination detection framework — error injection precision/recall, adversarial agent findings, and cross-reference consistency checks.

No eval results yet

Run crux evals run --suite=injection to generate baseline results.