AI & ML
Bad Apples in the Kitchen: How LLM Teams Detect, Correct, and Calibrate Against Their Misaligned Members
arXiv cs.MA0 views
arXiv:2611.22342v1 Announce Type: cross Abstract: What protects an LLM team when one teammate quietly goes wrong? Most multi-agent LLM work assumes cooperative partners and inspects chat between turns before acting. Real-time embodied cooperation does not allow this pause: chat runs alongside action, so a teammate whose words stay cooperative while behavior diverges can cause damage faster than the team can name what is happening. We study this case by injecting three theory-grounded misaligned agents into a custom four-player Overcooked environment, each stressing a different joint-action layer (speech-action mismatch, role invasion, and aggregate-progress disruption), and decompose robustness into three sequential gates of detection, correction, and calibration through a 3x4x2 design (n=720 episodes). Nearly all trolls are eventually named, but recovery follows only when attribution and situational diagnosis appear in the same utterance; direct source-attribution prompting improves the hardest case only by degrading the troll-free baseline by 37.5%. Robustness, then, depends less on detection alone than on whether team chat combines attribution with diagnosis in a single utterance, suggesting that sensitivity to misalignment should be targeted rather than uniformly raised.
Source:https://arxiv.org/abs/2611.22342PreviousMulti-Agent Memory Inconsistency: Where Factual Conflicts Disappear in LLM AgentsNextEquilibrium decisions under partial observation: dynamical error bounds and information requirements
Related posts
config init container for Nsigth Command
0dev.to Docker

Anthropic IPO is set to redraw VC power rankings, with lesser-known firms like Spark Capital poised for outsized returns compared with larger, blue-chip rivals (George Hammond/Financial Times)
0Techmeme

A Gallup survey of 15,925 people aged 15 to 22 across 16 countries finds that more than 33% rely on AI for companionship, rising to 50% among daily AI users (Jason Parham/Wired)
0Techmeme

London-based Payhawk, which offers enterprise expense and procurement management tools, raised a $280M Series C led by Goldman Sachs at a $2B valuation (Yazhou Sun/Bloomberg)
0Techmeme
Introduction to CUDA Beyond the Syntax
0Hacker News New
Physics AI: The illusion of ground truth
0Hacker News New