slashreboot 发布 Athena-Class Model A 实验模型
原标题：slashreboot/athena-class-model-a
AI 摘要
slashreboot 在 Hugging Face 发布实验性研究模型 Athena-Class Model A，基于 Gemma 4 31B Instruct 进行 LoRA 微调并合并量化为 Q8_0。该模型主打持久身份、长上下文一致性与无需系统提示的自主智能体行为，在 AIME 2026 上取得 93.33%（Consensus@4）、GPQA Diamond 85.86%。作者强调其为研究模型而非生产级系统，存在幻觉与量化质量下降等限制。
正文节选
--- license: apache-2.0 base_model: - google/gemma-4-31B-it - unsloth/gemma-4-31B-it base_model_relation: finetune language: - en tags: - Image-Text-to-Text - Transformers - gguf - gemma4 - long-context - self-modeling ---  # Athena-Class Model A **Experimental research release** Athena-class Model A is a specialized fine-tune of Gemma 4 31B focused on persistent identity, substrate-native self-modeling, independent research and agen