Back to AI News
arxiv

Live GPU Hidden State Acquisition Enables Request-Level Attribution in LLMs

Researchers present a method to extract per-layer residual-stream activations from a live GPU process using a non-modifying debugger, without prior model knowledge. The approach is validated on 33 open-source decoder-only models, achieving byte-exact recoverycars. Linear probes on recovered states distinguish refusal from compliance with AUROC 0.88, enabling forensic analysis of deployed LLM behavior.

AI Brief

Provides a concrete, novel technique for live AI system forensics with validated results, useful for security and compliance analysis.

arXiv:2611.18272v1 Announce Type: cross Abstract: Existing memory forensics for machine learning systems typically targets static artifacts, such as model weights or prompt logs, rather than the transient computational states of a live inference process. This gap limits an examiner's ability to analyze the specific behavior of a running large language model (LLM) service that cannot be restarted or modified. We present a method for acquiring per-layer residual-stream activations from a live GPU process using a non-modifying debugger, without requiring prior knowledge of the model's architecture parameters or weights. The approach employs a general six-stage process with platform-specific adapters, utilizing either a PyTorch allocator clock or a kernel-launch clock to identify and read tensor allocations from video memory. We validate this acquisition on 33 open-source decoder-only models across nine architecture classes, demonstrating that the recovered bytes are byte-exact and usable for downstream analysis. As a use case, we apply linear probes trained on standard framework activations to the recovered states, successfully distinguishing refusal from compliance with an AUROC of 0.88 on held-out prompts. This work establishes that live, observational access to intermediate hidden states is feasible on named inference platforms, providing a new source of evidence for examining the operational behavior of deployed AI systems.
Be the first to rate this article.

Comments & Questions (0)

Captcha
Please be respectful — let's keep the conversation friendly.

No comments yet

Be the first to comment!