conftrace_

token confidence

2 papers

Explore in graph

Co-occurring keywords · papers mentioning both

sequence lock (1) layer representation (1) large language model (1) real-time detection (1) backdoor detection (1) sliding window (1) representation steering (1) latent representation (1) machine unlearning (1) model unlearning (1)

Papers

ConfGuard: A Simple and Effective Backdoor Detection for Large Language Models AAAI 2026 On Effects of Steering Latent Representation for Large Language Model Unlearning AAAI 2025