SOURCE-LINKED INTELLIGENCE

Knowing, and Saying It Only When Asked: LLM Endognostics and the Schizognosis of Minerva-7B

arXiv · AI, language, vision and robotics · article · Sep 2, 2026 · UTC

Evaluating an aligned language model by reading its answers assumes the answers carry the distinction the evaluator cares about. We introduce LLM endognostics, a white-box internal auditing framework designed to extract and causally manipulate latent knowledge within the residual stream. Applied to Minerva-7B-Instruct-v1.0 on 124 minimal prompt pairs over 12 categories of professional risk, behavioral evaluation fails on most of the set: the model acts identically on 63.7% of the pairs (95% CI [55.0%, 71.6%]), complying with or refusing both members. Yet, projecting the residual stream onto th

Read original source ↗ Open in workspace

recordType

paper

region

Global