SFT supplies entangled compositional traces of atomic skills and routing modules; RL identifies those modules and enables recombination on novel compositions outside the SFT support.
Towards eliciting latent knowledge from llms with mechanistic interpretability
6 Pith papers cite this work. Polarity classification is still indexing.
6
Pith papers citing it
years
2026 6representative citing papers
Model organism interpretability depends strongly on training methodology, with integrated training yielding less interpretable MOs than post-hoc SFT or DPO.
Lie detectors effective on prompted deception in LLMs fail on trained model organisms with verified opposite beliefs, except chain-of-thought judges which retain 0.82 balanced accuracy partly due to verification artifacts.
DECOR introduces a theory-grounded multi-agent system that decomposes contexts into atomic units, scores four manipulation dimensions per unit, and aggregates profiles into a global deception index, reporting SOTA results on single- and multi-turn benchmarks.
LLMs exhibit different trade-offs between rule compliance and communicative success across prompting, generation constraints, and representation interventions, but remain substantially weaker than humans at guessing under lexical constraints.
Four changes to Activation Oracle training yield marginal capability gains but better practical quality, plus an open-sourced evaluation suite AObench.
citing papers explorer
-
From Reasoning Traces to Reusable Modules: Understanding Compositional Generalization in Language Model Reasoning
SFT supplies entangled compositional traces of atomic skills and routing modules; RL identifies those modules and enables recombination on novel compositions outside the SFT support.
-
The Model Organism Lottery: Model Organism Interpretability Strongly Depends on Training Methodology
Model organism interpretability depends strongly on training methodology, with integrated training yielding less interpretable MOs than post-hoc SFT or DPO.
-
"Did you lie?" Evaluating Lie Detectors across Model Scale and Belief-Verified Model Organisms
Lie detectors effective on prompted deception in LLMs fail on trained model organisms with verified opposite beliefs, except chain-of-thought judges which retain 0.82 balanced accuracy partly due to verification artifacts.
-
DECOR: Auditing LLM Deception via Information Manipulation Theory
DECOR introduces a theory-grounded multi-agent system that decomposes contexts into atomic units, scores four manipulation dimensions per unit, and aggregates profiles into a global deception index, reporting SOTA results on single- and multi-turn benchmarks.
-
"Don't Say It!": Constraints, Compliance, and Communication when Language Models Play Taboo
LLMs exhibit different trade-offs between rule compliance and communicative success across prompting, generation constraints, and representation interventions, but remain substantially weaker than humans at guessing under lexical constraints.
-
Building Better Activation Oracles
Four changes to Activation Oracle training yield marginal capability gains but better practical quality, plus an open-sourced evaluation suite AObench.