https://derekwang2002.github.io/ai-intelligence/en/projects/astabrief/ Material changes in AI models, agents, engineering and infrastructure, with primary sources and recommendations. en Mon, 02 Nov 2026 16:56:00 GMT https://derekwang2002.github.io/ai-intelligence/en/events/ev-20261002-01/ ev-20261002-01-published Fri, 02 Oct 2026 16:00:56 GMT New TRIAL <p>Ai2 released AstaBrief-8B, an Apache-2.0 model that turns a research question plus retrieved scientific excerpts into a cited report in one pass. It starts from Qwen3-8B, uses supervised fine-tuning and offline DPO, and ships with weights, training datasets and a local PDF example. On the 100-question ScholarQA-CS2 test set, Ai2 reports stronger answer-and-citation scores than its SFT checkpoint and similar aggregate quality to slower multi-step systems.</p><p><strong>Why read: </strong>AstaBrief-8B&apos;s public weights and local workflow make citation quality independently testable.</p><p>Recommendation: TRIAL</p><p>Sources: <a href="https://allenai.org/blog/astabrief">Ai2 introduces AstaBrief</a> · <a href="https://huggingface.co/allenai/AstaBrief_8B">AstaBrief-8B model card</a> · <a href="https://github.com/allenai/ai2-scholarqa-lib/tree/main/api/scholarqa/lite">AstaBrief local report-generation example</a></p>