Deep Reasoning Translation (DRT) Project

✨✨Latest Papers and Benchmarks in Reasoning with Foundation Models

RewardBench: the first evaluation tool for reward models.

A resource repository for representation engineering in large language models

A resource repository for machine unlearning in large language models

Accelerating your LLM training to full speed! Made with ❤️ by ServiceNow Research

🦦 Otter, a multi-modal model based on OpenFlamingo (open-sourced version of DeepMind's Flamingo), trained on MIMIC-IT and showcasing improved instruction-following and in-context learning ability.

[NeurIPS2023] Official implementation of the paper "Large Language Models are Visual Reasoning Coordinators"

[ CVPR 2024 ] Implementation for "GPT-4V(ision) is a Human-Aligned Evaluator for Text-to-3D Generation"

Memory optimization and training recipes to extrapolate language models' context length to 1 million tokens, with minimal hardware.

[ACL2025] Unsolvable Problem Detection: Robust Understanding Evaluation for Large Multimodal Models

[CVPR2025 Highlight] Insight-V: Exploring Long-Chain Visual Reasoning with Multimodal Large Language Models

Source code for the paper "Eraser: Jailbreaking defense in large language models via unlearning harmful knowledge".

The fastai book, published as Jupyter Notebooks

Convert TensorFlow, Keras, Tensorflow.js and Tflite models to ONNX

[ICLR 2023] Guess the Instruction! Flipped Learning Makes Language Models Stronger Zero-Shot Learners

[NeurIPS 2023] Tree of Thoughts: Deliberate Problem Solving with Large Language Models

PyTorch implementations of deep reinforcement learning algorithms and environments

First token cutoff sampling inference example

Lamorel is a Python library designed for RL practitioners eager to use Large Language Models (LLMs).

A simple and well styled PPO implementation. Based on my Medium series: https://medium.com/@eyyu/coding-ppo-from-scratch-with-pytorch-part-1-4-613dfc1b14c8.

Minimal implementation of clipped objective Proximal Policy Optimization (PPO) in PyTorch

PyTorch version of Stable Baselines, reliable implementations of reinforcement learning algorithms.

An implemtation of Everyting of Thoughts (XoT).

[ICML 2024] Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference

[ICML 2025] M-STAR (Multimodal Self-Evolving TrAining for Reasoning) Project. Diving into Self-Evolving Training for Multimodal Reasoning

B-STAR: Monitoring and Balancing Exploration and Exploitation in Self-Taught Reasoners

[NeurIPS'24] Official code for *🎯DART-Math: Difficulty-Aware Rejection Tuning for Mathematical Problem-Solving*

A repo for distributed training of language models with Reinforcement Learning via Human Feedback (RLHF)

Semantic cache for LLMs. Fully integrated with LangChain and llama_index.