Activation explainer for deception auditing in reasoning LLMs using hidden state decoding.

Amortized causal discovery model trained on synthetic data to map datasets to causal graphs.

Research paper explaining how Transformers learn modular multiplication using multiplicative character transforms.

A community-built, local-first alternative to generative design tools with multi-model support.

Multi-user white-box watermarking for contribution tracing in model derivation chains

A Rust image deconvolution and restoration crate.

An agentic system for autonomous climate science research and data analysis.

A collection of local-first browser tools for PDF, image, and dev tasks.

A framework for computer-using agents to accelerate repeated tasks via state-machine compilation.

Evaluation environment for measuring self-evolving LLM agent harness updates.

A customizable harness used to analyze agent trajectories and minimize the intent-execution gap.

Open-weight model used for private forensic analysis after commercial APIs blocked security queries.

A knowledge-guided multimodal perception and reasoning framework for autonomous medical systems.

Developer tool for agentic coding that provides day-one integration for GLM-5.2.

Hybrid NATS-MQTT orchestration platform for Edge Multi-Agent Systems.

A reactive notebook for Haskell allowing for interactive development and visualization.

Misspecification-Aware Simulation-Based Inference via Side-Channel Guidance

Subquadratic sequence modeling for efficient AI inference.

A benchmark for computer-use agents focusing on scientific instrument control via web-based simulators.

A multi-domain benchmark for demographic disparity in LLM agent actions.