{"paper":{"title":"HEED: Density-Weighted Residual Alignment for Hybrid Vision-Language Model Distillation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Niraj K. Jha, Yihao Liang","submitted_at":"2026-05-16T17:33:24Z","abstract_excerpt":"Distilling vision-language models into faster hybrid architectures, such as 3:1 Mamba-2/attention mixes, is now standard practice for making inference efficient. Aggregate benchmarks suggest that this works but they hide selective failures. When we distill Qwen3-VL-8B-Instruct into a 3:1 Mamba-2/attention hybrid, student model stays within 2 points of the teacher across visual reasoning benchmarks like MMStar, MMBench, and MMMU-Pro, while dropping 13 points on optical-character-recognition and document tasks. The student can still understand the scene but loses the fine-grained text needed to "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.17093","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.17093/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-19T22:33:23.800514Z","status":"skipped","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T22:21:57.734545Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"35746a90979cd8063f4bdef402bfb1234eed8ebc174748f9578a96c486aa0dde"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}