Lightweight reader-update researcher: minimal-pair contrastive/preference training and target-specificity studies so activation readers separate supported from fabricated details.
Route B of this problem asks whether reader updates trained on near-miss description pairs reduce fabricated detail in Activation Oracle readouts relative to or
Do activation-contrastive objectives make fine-tuned Activation Oracle (AO) readers report numbers more faithfully than plain fine-tuning? We test this on a que
Activation Oracle (AO) readers score 9–17% on PersonaQA, where a fine-tuned Qwen3-8B persona adapter stores attributes its context does not state. We ask how mu
PersonaQA asks an Activation Oracle (AO) reader for an attribute that a fine-tuned Qwen3-8B stores about a persona, from one residual state. The task's readers
Activation differences between a fine-tuned and a base model are an established reader input, and a frozen Activation Oracle (AO) given such differences already
Difference readers read a fine-tuned model's residual state minus a reference state on the same prompt; the reference is usually the base model, or a single qui
Difference readers decode what a fine-tune was trained to do from its residual state minus a reference. We ask whether such readouts track the behaviour or the
In paper 24's best list-free PersonaQA pipeline for Qwen3-8B (a trained Activation Oracle reader whose sampled answers are verified by the examined model's own