·
AI & ML interests
Post-training, model behavior and transparent evaluation.
Recent Activity
Organizations
published an article about 19 hours ago view article Five experiments inside Flint and Qwen Open
alphabell
• published an article about 23 hours ago view article One tool call, two input contracts: inspecting DeepSeek prompts
alphabell-research
• published an article about 24 hours ago view article Draw MOSS-Image layouts without doing coordinate arithmetic
alphabell
• view article alphaBell Qwen Open: 27B post-training, with the tradeoffs measured