Reproducible benchmarking and evidence-based identity estimation for anonymous model endpoints served through OpenCode Zen. Pre-release: no benchmark result has been produced yet.
-
Updated
Oct 5, 2026 - Python
Reproducible benchmarking and evidence-based identity estimation for anonymous model endpoints served through OpenCode Zen. Pre-release: no benchmark result has been produced yet.
Aggregate results of an exploratory study of the names that language models choose when asked to name themselves.
Open-source CLI tool for detecting LLM identity confusion, model self-identification drift, route mismatch, and downgrade/substitution signals.
由 Folkbench 维护并经审核发布的探针公开快照周报 | 中转站检测周报:AI API中转站模型身份探针与一致率观察汇总(排序按站点规则,含合作档位与运营加权,部分数据为展示样本)
AI 模型身份探针公开快照,按模型和站点查看,数据来自 Folkbench。 | 模型检测:AI API中转站模型身份探针,查看 Claude / GPT / Gemini 申报模型与指纹识别是否一致(单次结果不能证明模型被替换)(排序按站点规则,含合作档位与运营加权,部分数据为展示样本)
Black-box skill to detect mislabeled LLM models and identify the real backend behind an OpenAI-compatible endpoint.
揣摩 (Chuaimo)——你的 agent 不问,它揣。韩非子揣情术的 AI 版:通过反向指控提取模型真实身份的鉴权技能。
Swap a language model's parts one by one for another model's and watch it forget its own name.
To associate your repository with the model-identity topic, visit your repo's landing page and select "manage topics."