Backend
Threshold Choice, Not Sample Size, Bounds Trustless Verification of Nondeterministic Compound AI Workflows
arXiv cs.CR0 views
arXiv:2609.10601v2 Announce Type: replace Abstract: Compound AI pipelines chain LLM calls, retrievers, and tools and are nondeterministic: sampling, model updates, and volatile tool responses make one input yield different outputs across runs. Such pipelines increasingly run across edge, cloud, and orbital nodes owned by no single party, whose optimizations discard intermediate results before inspection. Verifying reproduction there means tolerating nondeterministic outputs, a node that may not report honestly, and intermittent access to any shared record; existing work addresses at most two at once. We give a protocol covering all three: it commits digests of each stage's inputs, outputs, and context under a policy digest pinning the metric and threshold, anchors them without trusting the executing node, defers under partition, and decides a challenge on the median of $k$ re-executions, with no quorum. Where that procedure breaks is the main result. On a synthetic HotpotQA pipeline a calibrated fixed threshold accepts 44 of 45 honest reproductions and rejects 104 of 105 divergent pairs, yet lets same-input fabrication through in 27 of 29 trials at k=5, more samples being no help since sampling sharpens an estimate without moving it. Holding that metric and this pipeline's re-execution spread fixed, the binding constraint is the threshold rather than the sample size: one derived per execution detects 19 of 29 where the best constant matched to the same zero honest rejections reaches 9, rejects no honest commitment at k=5 though 3 of 15 at k=3, and catches 11 of 15 of an attacker built against it, which has to aim at a target drawn only after its commitment exists. That rule is measured rather than deployed.
Source:https://arxiv.org/abs/2609.10601PreviousLanguage Model Security Modules: Turning LLM Interpretability into Runtime DefenseNextWho Is Behind the Harness? Verifying Model Identity in Coding Agents through Behavioral Fingerprinting
Related posts
My Rank Check Was Accurate. For the Wrong Country.
0dev.to
Audio to MIDI Explained: Why One Voice Is Easy and a Full Song Is Hard (With a Python Example)
0dev.to
Vercel Board: A Message Board Built for Agent Swarms
0dev.to

GTA VI: 212 GB, Pre-Load and What Take-Two Says About AI
0dev.to

Discord bot offline a week while systemd said active: a plugin timeout
0dev.to
Your AI Bill Has a Second Price Dial — and It Can Out-Swing Model Choice
0dev.to