Chinese contentPEARL 方法抑制语言模型不良行为，训练后不易复发
研究者把模型“越训越坏”的问题做成了可测量的框架，还给出 PEARL 这个 GRPO 改进版，做对齐或安全微调的可以看看怎么压制不良行为不反弹。
#GPT-OSS
8 stories · 0 in 7 days · 2 in 30 days
Stories, products and papers tagged "GPT-OSS", newest first.
Nov 4
Oct 21
Chinese content博主评价 6.1-Sol：CoT 风格与 GPT-OSS 相近
一位博主把 6.1-Sol 和 GPT-OSS 的思维链逐条对比，结论是两者说话风格差不多，适合关心模型行为差异的人看看。
Oct 4
Aleph Alpha's Kolibri: 78B open-weight model built for German
A European lab reaching benchmark parity with Chinese and American open models in a bilingual German-English model suggests small-model training capability has spread well beyond the usual hubs. The Apache 2.0 license makes it directly usable for on-premise deployments, which matters for German public-sector and enterprise buyers with data-sovereignty requirements. Teams comparing open models should weigh its German-language edge against its weaker SWE-Bench Verified result.
Oct 1
Sep 30
EASE framework for self-evolving agents
EASE addresses a critical limitation in agent skill curation by solving cross-executor degradation through behavior adaptation. Developers working with multiple AI models can now use a single, more efficient curator that reduces inference tokens by 9.1-14.5% while maintaining performance across diverse executors, making it particularly valuable for organizations deploying multi-model systems.
Sep 28
Chinese content论文：自监督置信度微调让推理模型生成 token 减少 25%
只用600道题微调，不训练停止机制，就让Qwen、Gemma这些模型推理时少生成25%的token，思路挺有意思。
Aug 16
Aug 11
