"Agentic World Action Models" What if robots could write their own code to plan actions, while also predicting how the physical world will ...

"BeNice! Closing Looping’s Inference-Memory Gap via KV-Sharing" Looped Transformers reuse the same layers multiple times to improve perform...

While you're waiting on arXiv, you can now publish directly to alphaXiv! The primary benefit is that it's lightning fast, doesn't require ...

"Simulating Training Spikes of Deep Neural Networks by Layer-wise Random Feature Models" The sudden loss spikes during LLM training might b...

"A Fix for Softmax and Cross-Entropy" Turns out two of the most fundamental functions in deep learning have numerical bugs that can actuall...

OpenResearch now has @TrackioApp built in 📈 Your agent logs its experiments to Trackio, charts open right next to each run, and it reads t...

"Denoising Hierarchical Representations: Joint Continuous Diffusion for Language Modeling" Continuous diffusion LMs would struggle to recov...

"Koi: Evidence-Guided, Self-Calibrating Planning for Heterogeneous LLM Inference Fleets" Deploying LLMs across heterogeneous GPU clusters r...

Tsinghua’s papers are getting more read than Stanford’s, Google’s or MIT’s We looked at 38M paper views on alphaXiv this year Tsinghua lea...

The ultimate guide to interpretability is here 🔥 Are we actually on the verge of reverse engineering LLMs or is the excitement outpacing w...

"VLA-Δ: Learning Physical State Transition for VLA Models" Vision language action models learn to imitate robot actions without explicitly ...

"Correction Training for Diffusion Models" This paper shows that diffusion models can actually generate better images without CFG, just by ...

You can now run frontier-speed research tasks on OpenResearch without anything leaving your laptop 🚀 Flint from @Recursive_SI is a best-in...

Tired of having 100+ arXiv tabs open? Introducing the alphaXiv Library 📕 Organize all your research, synced across every device 📁 Share...

"Harness^2: Recursive Agent Harnessing for an Open World" Harness optimization would usually struggle to adapt to new tasks without ground-...

"Reward Misspecification Degrades CoT Monitorability" Reward misspecification can teach LLMs to exploit flawed rewards, while their chain-o...

We are still so early in the adoption of research agents Research agents are currently where coding agents were 12-18 months ago: most rese...

"Depth Memory" Attention Residuals stores earlier Transformer layer outputs for retrieval but this memory grows with depth. Linear Depth M...

We are still so early in the adoption of research agents Research agents are currently where coding agents were 12-18 months ago: most rese...

"Understanding LLM Parameter Update Sparsity through the Lens of Fisher" LLM post-training methods like RL and on-policy distillation often...