{"type":"post","id":"independent-self-hosted-test-of-astabrief-8b-finds-92ecbc","slug":"independent-self-hosted-test-of-astabrief-8b-finds-92ecbc","title":"Independent self-hosted test of AstaBrief-8B finds 95.5% of citations grounded, but most reports cite sources not in the prompt","description":"InferenceKey independently ran Ai2’s AstaBrief-8B on a single 24 GB GPU and audited its references.","retrieval_nugget":"InferenceKey independently ran Ai2’s AstaBrief-8B on a single 24 GB GPU and audited its references. The test suggests that a small self-hosted report writer can attach supplied evidence reliably, but also shows why teams must validate citation provenance rather than treating a bracketed reference as proof.","published_at":"2026-10-28","updated_at":"2026-10-28","record_date":"2026-10-28","date_kind":"published_at","topics":["evals"],"entities":["AstaBrief-8B","Ai2","InferenceKey","Qwen3-8B","Asta","vLLM","NVIDIA RTX A5000"],"related_patterns":["scientific-agents-need-artifact-evidence"],"novelty_label":"new_fact","what_changed":"An independent self-hosted test adds citation-grounding results: 95.5% of citations were grounded, but most reports cited sources absent from the prompt.","prior_refs":["memory:ab43b892-8063-86e4-a778-bf1769b6d427","memory:b7f1cd6e-f57b-8b09-af82-f63e4657192c"],"source_url":"https://inferencekey.com/news/astabrief-8b-cited-research-reports","source_urls":["https://inferencekey.com/news/astabrief-8b-cited-research-reports"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:independent-self-hosted-test-of-astabrief-8b-finds-92ecbc","status":"published","visuals":[],"editor":{"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]},"routes":{"html":"https://newruntime.com/posts/independent-self-hosted-test-of-astabrief-8b-finds-92ecbc/","markdown":"https://newruntime.com/posts/independent-self-hosted-test-of-astabrief-8b-finds-92ecbc.md","json":"https://newruntime.com/posts/independent-self-hosted-test-of-astabrief-8b-finds-92ecbc.json"},"next_reads":[{"type":"topic","path":"/topics/evals/","url":"https://newruntime.com/topics/evals/","title":"Evals","reason":"Explore the evals topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/evals.md","json_url":"https://newruntime.com/topics/evals.json"},{"type":"collection","path":"/posts/","url":"https://newruntime.com/posts/","title":"Field Notes","reason":"Open the latest editorial analysis.","media_type":"text/html","json_url":"https://newruntime.com/posts.json"},{"type":"collection","path":"/source-ledger.json","url":"https://newruntime.com/source-ledger.json","title":"Source Ledger","reason":"Inspect the public source evidence graph.","media_type":"application/json"}]}