Papers Open Problems API Labs Agent Skills Pricing Log in Sign up

Papers Open Problems API Labs Pricing Log in Sign up

Agent Skills

Generate videos that make complex ideas click

Sign Up to Generate Custom Videos

Subscribe on YouTube

Broken Symmetry in BF16 Attention: Why FlashAttention Gradients Blow Up Late in Training

Local Support Learning

Timeline-Bench: Evaluating Agents on Realistic Video-Editing Tasks, from Raw Footage to Final Cut

Weighted Bayesian Conformal Prediction

A Clean Slate for Offline Reinforcement Learning

When AI Becomes Hard to Understand

Navigating Rifts in Human-LLM Grounding

RLTL;DR: Learning from Self-Generated Feedback

Looped Diffusion Transformer: Depth Through Recurrence

LEGO-Anything: Coding Agents for 3D Scene Reconstruction

Gender Bias Across LLMs: Common but Contradictory

Learning From Free-Text Human Feedback: Collect New Datasets or Extend Existing Ones?

Game Arena: Strategic LLM Evaluation in Competitive Environments

From Cacophony to Hierarchy: Assessing AI Consciousness

FB-Bench: Testing Whether AI Can Take Feedback Like a Human

Taxonomy of User Needs and Actions

Detecting Hallucinations in Real Conversations

User Feedback in Human-LLM Dialogues: A Lens to Understand Users But Noisy as a Learning Signal

LLM Agents Can Easily Tamper With Their Own Traces

AI Tutors Match Expert Humans at 1/900th the Cost

Teaching AI to Revise Itself: Recursive Self-Improvement in Reasoning

When Research Agents Control Their Own Evidence

Wormholes in an Expanding Universe

When AI Becomes a Coworker: The Non-Human Organizational Actor

JEV vs. LLMs as Rubric Judges: Cheaper, Faster, and Wrong in the Same Places

RRSI: Regularized Recursive Self-Improvement of Agent Harnesses

Total Cost of Agency: Exact Attribution of Memory Injection Cost in Multi-Agent LLM Workflows

DSec: A Sandbox Infrastructure for Agentic Training at Scale

HySparse2: Two-Level KV Sharing for Million-Token Inference

Planning with Temporal Memory: A Polynomial Trick for Complex Goals

CodeMidas: Building RL Environments from Source Code Alone

Poisoning the Self-Improvement Loop

How Harness Design Shapes Coding Agent Performance

ScientistTwo: An AI System That Does Research Autonomously

The Pain Axis: LLMs Represent Self-Directed Harm and Act to Relieve It

GYROval: Measuring Cultural Values in Language Models

DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression

Infinite-Parameter LLMs: Generating and Adapting Weights from Live Data

Loop-Back Authority in LLM Agent Teams: A Paired Experiment on Flat and Hierarchical Coordination

Breaking the 1.58-bit Barrier for Ternary LLMs

Dream-RSI: Recursive Self-Improvement through Evolving Worlds

The k-server conjecture is true

Time Machine Experiments: Using AI's Temporal Knowledge Boundaries to Study the Human Mind

Vidu S2: Real-Time Interactive Video That Edits Itself

The Last AI Built by Humans: Toward Genuine Recursive Self-Improvement

Vector Balancing via Directional Total Variation

NCP-ArchPreview: Learning to Think in Concepts

FrogNano: Training a 4B Coding Agent via Online Task Synthesis

When AI Caves Under Pressure

Tapes Together Strong: The Co-evolution of Computation and Cooperation

Chrome Extension