Mllms
Untracked preview. This is not a tracked Signal Bureau entity — the page is generated on request from live coverage search, is not indexed, and carries no editorial curation. Tracked entities live on the Wire.
If Mllms is on your watch list: what changed in the coverage, what the reporting asserts, and what the crowd is pricing.
What’s changed
- Seeing vs. Believing: Evaluating the Language Bias of Open-Source MLLMs in Counter-Intuitive Scenes · arXiv cs.AI · 2026-08-26
- Teaching MLLMs to Say No: Generalized Referring Expression Comprehension via Refusal Calibrated GRPO · arXiv cs.AI · 2026-08-06
- LongChart VQA: A Comprehensive Benchmark for MLLMs with Complex Multi-Chart Reasoning · ArXiv cs.AI · 2026-08-05
- Evidence-Grounded Trustworthy Multimodal Reasoning and Evaluation Benchmark in Complex Urban Scenes · ArXiv cs.AI · 2026-08-28
- AffectOmni: RL-Verifiable People-Centric Grounded Affective Reasoning for Social and Art-Related Scenes · Source · 2026-08-28
- MolEmb: Multimodal Large Language Models Can Be Strong Molecular Embedding Models · Source · 2026-08-26
What the reporting says
It shows up alongside Large Language Models (17), Abstract (17), Computer Science (17), View Pdf Html (16), Citations Loading Bibliographic (15), Connected Papers (14), Explorer (14), Litmaps (14).
Two steps out: AI (through Large Language Models, 17 shared stories on the weaker link); Catalyzex (through Abstract, 17 shared stories on the weaker link); Dagshub (through Abstract, 17 shared stories on the weaker link); Gotit Pub (through Abstract, 17 shared stories on the weaker link).
Tracked since Jul 2, with coverage on 11 separate days — concentrated in Artificial Intelligence and Auto.