Flash-dLLM is a training-free inference acceleration framework for diffusion large language models that identifies GPU memory I/O, rather than compute,…
SiliconBench benchmarks nine LLM serving engines on Apple Silicon unified-memory desktops across speed, memory headroom, and output fidelity rather than…
Agensh is a multi-agent harness that removes the central-orchestrator bottleneck limiting existing multi-agent frameworks by having concurrent workers self-organize —…
The paper presents AIDE^2, a system in which an AI research agent recursively rewrites and benchmarks its own code, keeping…
Bellman Policy Optimization (BPO) is a critic-free reinforcement learning method for LLM reasoning that reformulates Policy Mirror Descent as a…
StableVQ addresses training instability in vector-quantized visual tokenizers by arguing the root cause is entanglement between encoder-decoder and codebook training,…
JEV-as-a-Judge is a decision-only LLM evaluation model designed as a cheap first-pass filter, tested against sixteen generative and reward-model judges…
This paper introduces mutation analysis to measure whether GPU-kernel benchmark correctness checkers — which now feed leaderboards and RL reward…
StepFun unveiled Step 5 Preview, a proprietary 600-billion-parameter sparse mixture-of-experts model aimed at long-horizon agent workloads such as software engineering,…
President Donald Trump announced plans to create a new ‘AI Force,’ modeled on the Space Force, and to appoint an…
Anthropic named Accenture as its first embedded evaluator, tasked with observing model development from inside the company as part of…
An unsealed court filing in The New York Times’ copyright lawsuit against Microsoft and OpenAI revealed that a Microsoft director…
