LLM Digest

Story

arxiv_cs_cl · Oct 30, 2026 · paper

Source brief

When Steering Shifts Discourse: Identifying and Intervening on Discourse-Level Changes due to Activation Steering in LLMs

arxiv.orgOct 30, 2026
original source linked

In brief

Activation steering is an inference-time technique used to control or modify the behavior, tone, or safety properties of a large language model (LLM) without retraining the model's weights or editing its input prompt,...

Read the original at arxiv.org →Open in live feed Read that day’s brief

When Steering Shifts Discourse: Identifying and Intervening on Discourse-Level Changes due to Activation Steering in LLMs

Earlier in this thread 4 items

Accept Changes, Reject Attacks: Robust Memory Management for LLM Agents

Composing and Steering Diffusion Models with Generalized Feynman-Kac Correctors

How we test LLM serving changes for coding agents, two hours of GPUs per change

CANS: Capability-Preserving Neuron Steering for Large Language Models