… To evaluate these capabilities, we build a multimodal presentation benchmark covering single presentation , discussion , and interaction scenarios, with task-specific evaluation criteria for content quality, media relevance, dynamic media use, dialogue naturalness , and interaction grounding . …
… We achieve competitive performance with state-of-the-art detectors AUROC 0.927 while natively providing annotations that explain the basis for the detector's decision. …
… While limited by self-reported ground truth, these results demonstrate the benefits of a dedicated and complete symptom interview compared to a user-guided symptom discussion, which is the default of most consumer LLMs. …
… On average, dMoE reduces the number of uniquely activated experts from 69.5 to 14.6 while retaining 99.11% of the original performance. …
… For in-domain performance, DeBERTa-large achieved the highest F1 score F1=0.844 as well as for the X Twitter out-of-domain test ACC=0.864. …
… Finally, we sincerely thank the community for its attention, discussion, and promotion of Cola DLM. …
… Feedback and discussions are welcome. …
… Existing KV eviction methods reduce this cost but typically degrade performance relative to full-cache inference. …
… View arXiv page View PDF GitHub 35 Add to collection Community Welcome everyone to join the discussion! …
… That would be a pretty cool move :- ModernBERT's success is delightful, and its performance in downstream tasks is astonishing. …