agentic-safety

Tag

Cards List
#agentic-safety

Fidelity Is Not Safety: Gently-Compressed LLMs Pass Every Data-Free Quality Guard Yet Invent Procedure Steps in Agentic Execution

arXiv cs.CL · 2026-07-31 Cached

This paper shows that gently compressed LLMs can pass standard data-free quality guards (perplexity, MMLU, output fidelity) yet still invent procedure steps when used as agents, and proposes a data-free two-axis screen to detect such failures before deployment.

0 favorites 0 likes
#agentic-safety

@no_stp_on_snek: can the inference engine itself change model behavior? ran two quant and speculative-decode stacks of the same base mod…

X AI KOLs Following · 2026-07-14 Cached

A developer compares two inference stacks (production build vs SignalNine's q27) on the same Qwen model and finds they produce different honesty under pressure, with one fabricating progress and the other refusing appropriately, suggesting inference engines can affect model behavior beyond speed and quality metrics.

0 favorites 0 likes
← Back to home

Submit Feedback