Implement a reasoning LLM in PyTorch from scratch, step by step

Repo for AI Agents The Definitive Guide

Inference-time scaling for LLMs-as-a-judge.

Official repository of the spotlight ICML 2025 paper, PokeChamp: an Expert-level Minimax Language Agent.

Multi-Agent Verification: Scaling Test-Time Compute with Multiple Verifiers

[NeurIPS 2025 Spotlight] Official implementation of "Learning with Calibration: Exploring Test-Time Computing of Spatio-Temporal Forecasting"

Test-Time Memory Framework: Control Hallucinations in Foundation Models

Code for the paper "Specification Self-Correction: Mitigating In-Context Reward Hacking Through Test-Time Refinement"

Code for ICML 2025 How Do Large Language Monkeys Get Their Power (Laws)?

An experimental project using MCTS to refine LLM responses for better accuracy and decision-making.

ReasonForge — Interactive Test-Time Compute Laboratory. Live scaling curves, MCTS+PRM, Tree-of-Thoughts, Best-of-N verifier, Self-Refine reward-hacking — all on the Game-of-24 benchmark.

A Framework Enabling Web Agents to Master Workflows From Human Demonstration

Research pilot (dormant): steering test-time compute using answer-incoherence as a signal (GPQA/MMLU)

Can 3B + test-time compute match 70B on network troubleshooting?

Combinatorial group testing for verifier-cost-limited reasoning: localize defective reasoning steps from ~k*log(n/k) pooled verifier queries under an explicit asymmetric noise model (simulated headline; pre-alpha).

Hands-on demo mapping the accuracy-vs-token-cost Pareto front of LLM test-time scaling strategies: chain-of-thought, self-consistency, debate, and mixture-of-agents.

DTTC is a lightweight framework designed to enhance the reasoning capabilities of small language models (SLMs). By introducing a Dynamic Parameter Pool (DPP), Ambiguity Statement Mapping (ASM), and Time-enhanced Penalty Decoding (TPD), we significantly bridge the gap between 1.5B models and larger state-of-the-art LLMs.

TypeScript CLI + MCP server implementing OpenDeepThink (Zhou et al., arXiv:2605.15177). Independent reimplementation; not a fork.

Deliver predictive litigation modeling and outcome simulation with enterprise-grade legal analytics for high-stakes trial intelligence.