Tag
A user claims that an API fingerprint test revealed DeepSeek V4.1 Flash (ds-4.1-flash-1017) may be related to GPT-6 Astra, joking that DeepSeek's pricing lets you enjoy a GPT-6 level model.
Sentient Foundation 发布了免费开源的 Fingerprinting Leaderboard 基准测试平台及配套工具 mlprints，用于评估 AI 模型指纹（所有权标记）的可读性、抗移除能力和对模型性能的影响，其自家的 Perinucleus 方法也参与排名。
Sentient Foundation 介绍了 Fingerprinting Leaderboard，一个针对模型指纹技术的开放基准，用于在模型权重公开后验证其归属。该基准从模型质量保留、抗移除能力以及检测可靠性三个维度对各类指纹方法进行评分。
论文提出 ForgePrint（搜索-蒸馏框架），通过定向改写将一个模型的输出伪装成目标模型的指纹，使归因分类器误判来源模型，在 CNN/DM 上达到 70.2% 的目标成功率，说明文本归因不能等同于来源真实性。
A developer is creating a method to visually fingerprint AI models by analyzing weights and tensors using Random Matrix Theory, with data from QWEN2.5 models, and is interested in open-sourcing the tool.
A technical blog post presenting a reproducible pipeline to fingerprint whether an LLM was trained from scratch or derived from existing open weights, using architecture, tokenizer, and weight comparisons.
User @qqqqqf_ shared fingerprint test results for Fable5 and other models, finding that Fable5 has low similarity with most Anthropic models but very high similarity with kimi k2.7, qwen3-7max, glm5.2, etc., speculating possible data leakage or distillation. At the same time, they released an open source project ai-model-fingerprint for model identification and detection.