Qwen3.5-4B-Base
PulseAugur coverage of Qwen3.5-4B-Base — every cluster mentioning Qwen3.5-4B-Base across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
New research questions value of trajectory features for LLM attention routing
Researchers have evaluated the effectiveness of trajectory features for routing attention in the final layer of language models. Using frozen checkpoints of SmolLM3-3B-Base and Qwen3.5-4B-Base, they tested utility-super…
-
GrowMTP trains draft heads within RL loop, accelerating LLM training
Researchers have developed GrowMTP, a novel method that trains a draft head for speculative decoding entirely within the reinforcement learning (RL) loop. This approach eliminates the need for pre-training draft heads s…
-
AMD releases open Instella-MoE-16B LLM with 2.8B active parameters
AMD has released Instella-MoE-16B-A3B, an open-source Mixture-of-Experts language model. This model features 16 billion total parameters but only activates 2.8 billion per token, utilizing architectural innovations like…
-
AMD releases open-source Instella-MoE AI model trained on its GPUs
AMD has released Instella-MoE, a new open-source Mixture-of-Experts language model developed using their own GPUs and software. The model is available in various forms, including pre-trained, mid-trained, and fine-tuned…
-
New LKValues resource aligns LLMs with Sri Lankan societal values
Researchers have developed LKValues, a new resource suite designed to align large language models (LLMs) with the specific societal values of Sri Lanka. Existing benchmarks often reflect Western norms, failing to captur…
-
New PMC-InterCPT dataset enhances biomedical multimodal models
Researchers have developed PMC-InterCPT, a new dataset designed to improve multimodal models for biomedical applications. This dataset addresses limitations in existing image-text pairs by incorporating relevant surroun…