AINewsnow

When Keywords Drop but Classifiers Hold: Soft Refusals under KV Cache Compression

arXiv:2609.31678v1 Announce Type: new Abstract: KV cache compression is widely used for long context LLM inference under memory constraints, while deployed systems typically score refusals after generation with keyword filters or learned classifiers. Such monitors are intended to indicate whether a…

Read the full story at arXiv cs.LG ↗

Timeline · 1 report

  1. 2026-09-29 04:00 · arXiv cs.LG
    When Keywords Drop but Classifiers Hold: Soft Refusals under KV Cache Compression

More stories

  1. NVIDIA Open Agent Safety Platform: A Reference for Continuous In-Silicon Agent Monitoring — NVIDIA Technical Blog
  2. How we found 24 Android vulnerabilities using our open source AI security agent — GitHub Blog
  3. OpenAI Scraps Debut of Latest Astra Model Over Safety Risks — Bloomberg AI
  4. Introducing Claude Sonnet 5.5 on AWS — AWS Machine Learning Blog
  5. Heads of OpenAI and Anthropic called to face Senate inquiry after rogue agent incidents — The Guardian AI
  6. Meta launches enterprise AI business seeking to cash in on vast spending — Financial Times AI
  7. AMD agrees to acquire Fei-Fei Li's World Labs for $8.2B in an all-stock deal expected to close by year-end; Li will join AMD as EVP and chief scientist (Edward Ludlow/Bloomberg) — Techmeme
  8. Scoop: Anthropic's Dario Amodei to have White House dinner with Trump — Axios AI+

Get the daily brief of stories like this at 6:30 every morning →