Tracked topic
Qwen3 is an AI model family developed by Alibaba, released as a set of large language models for natural-language tasks.
…Across six reasoning tasks, Qwen3 models using VaSE with 4x KV cache compression yield higher average accuracies than SOTA selection method at the same sparsity, while outperforming the strongest eviction method by…
…Compatible with Qwen3, Qwen3.5, Llama3.1, etc. The target model runs in a multi-stage pipeline while a lightweight speculation head drafts tokens in parallel; drafts are verified against the base…
…On Qwen3-30B-A3B , MOPD outperforms Mix-RL , Cascade RL , Off-Policy Finetune , and Param-Merge baselines, inheriting nearly all of each teacher's capability. MOPD also enables parallel, independent development of…
…Built upon the Qwen3 language model (1.7B and 4B), BioMatrix is continually pretrained on 304.4 billion tokens spanning general and domain-specific text, sequence and structure views of molecules and…
…Across 14 MLLMs , even the strongest model reaches only 65.6% transition validity and 66.3% requirement coverage, and visual quality is no proxy for behavior (Qwen3.6-35B-A3B on Markdown…
…Evaluations on mathematical reasoning benchmarks using Qwen3 models reveal an emergent phenomenon : geometric latent reasoning induces substantially shorter generations without an explicit length objective. By replacing early explicit reasoning with continuous latent…
…Using a frozen Qwen3-4B solver, we log intervention outcomes and train recoverability-aware gates from serving-visible attempt state. On \mathfive, selective verification reaches 76.3\% accuracy, compared with 75.5…
…On ScreenSpot-Pro , it raises Qwen3-VL 4B/8B/30B-A3B from 55.5/52.7/53.7 to 63.4/65.8/67.0. Robustness analyses further show higher worst-view…
["qwen3","alibaba"]
…Specifically, RACES improves DeepSeek-R1-Distill-Qwen-14B by an average of 3.1 points (from 48.2 to 51.3) and boosts Qwen3-14B performance from 58.8 to 61.1…