阶跃星辰开源 LLM 标注与模型检查工具 onPanda
We’ve open-sourced onPanda 🐼 — the tool we use internally for LLM data annotation and model inspection.
阶跃星辰（StepFun）开源了内部用于 LLM 数据标注与模型检查的工具 onPanda，工作流为定位错误、修正 token、让模型继续生成。其数据标注中位耗时比人工后编辑低 52%，SFT 与偏好数据在同一流程完成，on-policy 保真度高（ΔPPL <1%），支持 token 级正负样本监督及图像、音频、视频的 agent 轨迹标注；模型检查可查看 token 概率与 top-k 备选、逐 token 引导解码，并在浏览器中探索 SVG 生成、Web 开发和 agent 任务。
查看原帖I spent two years building this interactive tool to let you steer LLMs and agents at the token level. Introducing onPanda — a web app for token visualization & control, model inspection, data annotation, and more. Try it online (works on mobile): https://t.co/SFIPmeTZ2T
正文
当前显示原文，中文全文译文尚未就绪。
We’ve open-sourced onPanda 🐼 — the tool we use internally for LLM data annotation and model inspection.
The workflow is simple: find an error, correct the token, and let the model continue.
✍️ Data annotation - 52% lower median annotation time vs. manual post-editing - SFT + preference data in one workflow, with high on-policy fidelity (ΔPPL <1% vs. the model’s resampling baseline) - Precise token-level supervision with paired positive/negative examples, plus agent-trajectory annotation across image, audio, and video
🔎 Model inspection and debugging - Inspect token probabilities and top-k alternatives, steer decoding token by token, and explore SVG generation, web development, and agent tasks directly in the browser.
Try it (mobile-friendly): https://t.co/57vAeJetWD