GUARD-IT performs machine unlearning in LLMs via inference-time gated activation redirection, matching or exceeding gradient-based baselines on TOFU and MUSE while preserving utility and working under quantization.
Title resolution pending
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
representative citing papers
Compositional interpretability defines explanations as commuting syntactic-semantic mapping pairs grounded in compositionality and minimum description length, with compressive refinement and a parsimony theorem guaranteeing concise human-aligned decompositions.
Activation Addition steers language models by adding contrastive activation vectors from prompt pairs to control high-level properties like sentiment and toxicity at inference time without training.
Natural-language descriptions generated and verified through generative models and digital twins capture the selectivity of most neurons in macaque V1 and V4.
LLMs perform in-context learning as trajectories through a structured low-dimensional conceptual belief space, with the structure visible in both behavior and internal representations and causally manipulable via interventions.
Stable personality vectors in LLMs function as intrinsic guardrails, with ablation increasing emergent misalignment above 40% and amplification reducing it below 3%, enabling zero-shot transfer from aligned to corrupted models.
Bilinear autoencoders decompose neural activations into low-rank quadratic forms to discover interpretable multi-dimensional manifolds, improving reconstruction in language models and challenging linear representation assumptions.
citing papers explorer
-
Inference-Time Machine Unlearning via Gated Activation Redirection
GUARD-IT performs machine unlearning in LLMs via inference-time gated activation redirection, matching or exceeding gradient-based baselines on TOFU and MUSE while preserving utility and working under quantization.
-
From Mechanistic to Compositional Interpretability
Compositional interpretability defines explanations as commuting syntactic-semantic mapping pairs grounded in compositionality and minimum description length, with compressive refinement and a parsimony theorem guaranteeing concise human-aligned decompositions.
-
Steering Language Models With Activation Engineering
Activation Addition steers language models by adding contrastive activation vectors from prompt pairs to control high-level properties like sentiment and toxicity at inference time without training.
-
Letting the neural code speak: Automated characterization of monkey visual neurons through human language
Natural-language descriptions generated and verified through generative models and digital twins capture the selectivity of most neurons in macaque V1 and V4.
-
Stories in Space: In-Context Learning Trajectories in Conceptual Belief Space
LLMs perform in-context learning as trajectories through a structured low-dimensional conceptual belief space, with the structure visible in both behavior and internal representations and causally manipulable via interventions.
-
Intrinsic Guardrails: How Semantic Geometry of Personality Interacts with Emergent Misalignment in LLMs
Stable personality vectors in LLMs function as intrinsic guardrails, with ablation increasing emergent misalignment above 40% and amplification reducing it below 3%, enabling zero-shot transfer from aligned to corrupted models.
-
Bilinear autoencoders find interpretable manifolds
Bilinear autoencoders decompose neural activations into low-rank quadratic forms to discover interpretable multi-dimensional manifolds, improving reconstruction in language models and challenging linear representation assumptions.