3

Speculative Probing: LLM Monitoring at Speculative-Decoding Cost

Real-time classification during language model inference is valuable for safety filtering, behavioral analysis, and model monitoring, …

PDF Cite Code
Deep-Research Agents Can Be Poisoned via User-Generated Content
Deep-research agents, i.e., systems that rely on multi-agent pipelines to iteratively retrieve, synthesize, and cite Web content in …
PDF Cite Code Tweet 404 Media Nature Le Monde NRC El País Corriere della Sera
How to Steal Reasoning Without Reasoning Traces
Many large language models (LLMs) use reasoning to generate responses but do not reveal their full reasoning traces (a.k.a. chains of …
PDF Cite Code Tweet