720K public decision-training samples released for building tiny in-house decider models
MaziyarPanahi · x · 2026-11-06
MaziyarPanahi published the "Jevmaxxing" collection on Hugging Face: 720,000 decision answers across 5 public datasets covering tool calls, agent traces, customer support, intent routing, and policy checks (e.g., AgentToolDecisions-180K, Jev-Policy-Verification-180K, Jev-Customer-Support-Decisions-130K, Jev-Agent-Trace-Audit-100K), totaling 1.8B tokens read. Every row includes a source label and Jev's answer so you can train your own tiny decider model and keep it in-house (e.g., in a hospital). The collection also includes ModernJEV-Decide-Preview (0.1B) and Qwen3.5-4B-Jev-Judge models.
Related event: Jev Releases 720K Decision Data for Training Lightweight Routing Models(3 posts)→
More from coding & agent
- 20 open-source AI agent observability tools worth knowing in 2026, categorized — MaryamMiradi · 2026-11-06
- Ampersand Claims 92.2% on SWE-bench Pro HARD-51 at 70% Lower Cost — niloofar_mire · 2026-11-06
- ServiceNow Brazil Release Touts 100x Gains with Autonomous Engineer — azavery · 2026-11-06
- LiteParse Integrates 1.3B MiniCPM-OCR, Beating Gemini Flash on ParseBench — llama_index · 2026-11-06
- Finix MCP Server lets AI agents create buyers, payment links, and search docs — modelcontextprotocol · 2026-11-06
- Agent Substrate ships an MCP server for fs/bash on actors with auto suspend-resume — rakyll · 2026-11-06