AI & ML
Game-Theoretic Data Allocation for Expert Specialization and Routing
arXiv cs.MA0 views
arXiv:2611.25167v1 Announce Type: cross Abstract: Can the allocation used to train experts also predict which expert will perform best on a prompt? Such a connection would make the intended division of labor inspectable before training and allow the allocation itself to guide inference. We introduce InflEns, a mixture of LoRA experts whose training weights and inference routes come from the same location game in a fixed prompt-trait space. Players choose positions to maximize their expected shares of the prompt distribution, without using expert losses. We utilize the game's kernel-general theoretical criterion for when co-located positions become locally unstable to separation. We evaluate InflEns on a corpus assembled from seven safety-related benchmarks, which demonstrates that the highest-share expert has the lowest held-out response negative log-likelihood (NLL) on approximately 90% of prompts across four kernels under a sharpened simplex representation, compared with approximately 15% after permuting allocations across prompts. We demonstrate that the method performs comparably to published learned gates without having to train an explicit gate on the experts. Because the router is fixed, we can trace harmful behavior to a single expert and retrain only that expert, which lowers HarmBench attack success from 0.51 to 0.02 while keeping instruction following. The game also gives us several controls over how the experts specialize, including the kernel choice and its parameters, the trait-space geometry, and the number of experts. Importantly, InflEns is a general framework that can be extended to any encoder and data distribution.
Source:https://arxiv.org/abs/2611.25167PreviousGridAnchor: Specialized Multi-Agent AI for Safeguarding Autonomous Energy Systems Against Cyber-Physical HallucinationsNextSelf-Evolving Multi-Agent Framework for Efficient Decision Making in Real-Time Strategy Scenarios
Related posts

So sicherst du einen MCP-Server mit codex sandbox ab
0dev.to

Perfiles de permisos de Codex: ejemplos de config.toml para ejecuciones seguras de agentes
0dev.to

GPTBot vs OAI-SearchBot: How to Configure robots.txt for AI Crawlers
0dev.to

ملفات تعريف أذونات Codex: أمثلة config.toml لتشغيل الوكلاء بأمان
0dev.to

Write a cover letter with AI and keep every fact yours: a nine-step loop
0dev.to
Reading project health off a download page — faster than star count
0dev.to