This work challenges the assumption that recurrent linear-attention layers require higher precision than other model components, showing that the Gated…
This paper addresses the memory bottleneck created by large key-value caches during extended LLM reasoning by proposing Random Attention, which…
This work tackles a performance bottleneck in attention computation on Blackwell GPUs’ 4-bit floating-point tensor cores, where softmax conversion and…
This paper addresses the difficulty of assigning credit across long, multi-step agent trajectories during reinforcement learning, where sparse end-of-episode rewards…
This paper addresses the scarcity of executable environments needed to train terminal-based coding agents by reconstructing usable environments directly from…
This paper addresses the inefficiency of storing and processing long conversational histories and documents by introducing a machine-facing compression interface…
This paper investigates on-policy distillation (OPD) by studying what happens when training uses minimal data, down to a single query,…
This paper addresses the slow, strictly sequential token generation inherent to autoregressive language models by introducing diffusion-augmented LLMs that define…
This paper addresses text-transformation tasks that are conceptually simple to describe but hard to implement with traditional rule-based code, while…
This paper addresses the lack of standardized interoperability between AI agents built on different frameworks and deployed across organizations. It…
The study introduces a benchmark evaluating whether LLMs can construct and iteratively refine their own agent execution harnesses rather than…
The paper presents DisCo, a research agent that distills operational knowledge from GitHub repositories into reusable AI skills using two…
