AI & ML
From Bytes to Behavior: Request-Level Attribution from Live GPU Hidden States
arXiv cs.CR0 views
arXiv:2611.18272v1 Announce Type: new 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.
Source:https://arxiv.org/abs/2611.18272PreviousForget the Lineage: Forward-Secure, Lineage-Complete Cryptographic Erasure for LLM-Agent MemoryNextPrivacyPR: A Dataset of Real-World Privacy Fixes for Benchmarking Privacy-Aware Code Review
Related posts
Building an AI Agent Backend with FastAPI: From User Request to AI Response
0dev.to AI

SoduWork Custom Model Setup: Connect an OpenAI-Compatible Gateway and Handle the 180-Second Stall
0dev.to AI
AI startup hiring snapshot, 16 November 2026: 6,481 open roles, half at four companies
0dev.to AI
EC2 已经停止,账单为什么还在变?用资源清单代替猜测
0dev.to AWS

Amazon Black Friday 2026:5 Key Metrics and a Step-by-Step Data Collection Guide
0dev.to AWS
How to Trim Audio for an Online Course Lesson — Free, in Your Browser
0dev.to WebDev