toolhorizon

provenance:github:v0lcanofish/toolhorizon

单卡 24GB 上研究长链路多工具 Agent 的 RL 训练稳定性 —— 以及工具使用行为本身的退化。Qwen2.5-1.5B + LoRA / τ-bench airline / 四臂消融。RL training stability for long-horizon multi-tool agents on a single RTX 4090D.

View Source ↗First seen todayNot yet hireable

PUBLIC HISTORY

First discoveredSep 21, 2026

IDENTITY

inferred

Identity inferred from code signals. No PROVENANCE.yml found.

Is this yours? Claim it →

METADATA

platformgithub

first seenSep 20, 2026

last updatedSep 20, 2026

last crawledtoday

version—

README BADGE

Add to your README:

![Provenance](https://getprovenance.dev/api/badge?id=provenance:github:v0lcanofish/toolhorizon)