3 papers
cs.AI2026
Decodability is Not Causality: Dissociating Probe Readouts from Behavioral Drivers via SAE Decomposition
Devesh Tiwari, Camille Davis, Shivank Sinha +3
Linear probes can decode safety-relevant concepts such as truthfulness from language-model activations, but probe accuracy may show only decodability, not that the features the pro…
cs.CL2026
The Diminishing Returns of Early-Exit Decoding in Modern LLMs
Rui Wei, Rui Du, Hanfei Yu +4
In Large Language Model (LLM) inference, early-exit refers to stopping computation at an intermediate layer once the prediction is sufficiently confident, thereby reducing latency…
cs.AI2026
RLHFless: Serverless Computing for Efficient RLHF
Rui Wei, Hanfei Yu, Shubham Jain +5
Reinforcement Learning from Human Feedback (RLHF) has been widely applied to Large Language Model (LLM) post-training to align model outputs with human preferences. Recent models,…