The paper introduces TASCO, a test-time adaptation framework that improves LLM reasoning by optimizing lightweight task-level prefixes while keeping the…
The paper proposes a unified four-coefficient parameterization for on-policy knowledge distillation that generalizes existing per-token gating methods by combining forward…
Researchers introduce AgentZip, a memory compression system for AI-agent sandboxes that exploits redundancy across sandbox instances and between sandboxes and…
COBRA-Skills is a framework that optimizes reusable skills for LLM agents by framing skill development as budgeted sequential optimization over…
The paper studies when to retrain an active-learning model from scratch versus fine-tune it, finding retraining wins early and fine-tuning…
The paper studies agent harnesses (system prompt, tools, execution hooks, context management) together with model fine-tuning, finding that naively imitating…
The authors train two small code models (a 100M-parameter Rainbow-Pony-100M and a fine-tuned Qwen2.5-Coder-0.5B) on Flutter/Dart code in both direct…
The team describes OreoLook, an open-source AI answer-engine (browser-agent-driven, provider-routed LLM inference) whose local caching stack runs entirely on commodity…
Researchers introduce EvoSafeHarness, a framework that automatically generates customized safety policies for LLM agents by jointly optimizing a natural-language policy…
Researchers present SenseNova-U1.5, an 8-billion-parameter multimodal model that performs visual understanding, reasoning and generation within a single unified architecture, eliminating…
Researchers introduce NCP-ArchPreview, a latent-space language model that extends standard autoregressive pretraining by adding a Next Concept Prediction objective alongside…
Researchers introduce Recursive Code World Models (RCWM), a framework for reconstructing detailed 3D environments from a single image by representing…
