CLAS dynamically adapts linear activation steering strengths to context, outperforming fixed-strength steering and matching or exceeding ReFT and LoRA on eleven benchmarks across four model families with limited labeled data.
Title resolution pending
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
verdicts
UNVERDICTED 3representative citing papers
CR-Net uses cross-layer low-rank residuals in a dual-path network plus specialized recomputation to outperform prior low-rank methods on 60M-7B model pre-training while using less compute and memory.
UNSEEN combines AR access control, LLM unlearning to suppress profiles, and agent guardrails to defend against AR-LLM social engineering attacks, tested in a 60-person user study with 360 conversations.
citing papers explorer
-
Contextual Linear Activation Steering of Language Models
CLAS dynamically adapts linear activation steering strengths to context, outperforming fixed-strength steering and matching or exceeding ReFT and LoRA on eleven benchmarks across four model families with limited labeled data.
-
CR-Net: Scaling Parameter-Efficient Training with Cross-Layer Low-Rank Structure
CR-Net uses cross-layer low-rank residuals in a dual-path network plus specialized recomputation to outperform prior low-rank methods on 60M-7B model pre-training while using less compute and memory.
-
UNSEEN: A Cross-Stack LLM Unlearning Defense against AR-LLM Social Engineering Attacks
UNSEEN combines AR access control, LLM unlearning to suppress profiles, and agent guardrails to defend against AR-LLM social engineering attacks, tested in a 60-person user study with 360 conversations.