Stars
When a model translates out of English and then back to English, how much meaning and voice does it keep? Each model does both steps (English → target language → English).
Daytona is a Secure and Elastic Infrastructure for Running AI-Generated Code
Benchmarking for Evaluating Web Experience (Core Web Vitals, etc)
Repository for "LoRA-Mini : Adaptation Matrices Decomposition and Selective Training (AAAI Workshop 2025)"
AOT-GPT: A JIT compiler that uses GPT to optimize Python → LLVM IR, then formally verifies correctness with Z3 before compiling to native code.
[ICML 2025] UDora: A Unified Red Teaming Framework against LLM Agents
[KDD 2024] "ImputeFormer: Low Rankness-Induced Transformers for Generalizable Spatiotemporal Imputation"
MACE-RL is a novel framework that learns to dynamically regulate its own exploration strategy. It integrates a curiosity module, informed by episodic memory of past experiences, with a meta-learnin…
The project investigates whether learned world models can effectively simulate and plan in discrete, logic-intensive environments.
This repository implements HCMRL (Hierarchical Compressed Memory for Resource-Limited RL), a novel approach to memory-augmented reinforcement learning that is specifically designed for resource-con…
SIDiffAgent autonomously manages prompt engineering, detects and corrects poor generations, and performs fine-grained artifact removal, yielding more reliable and consistent outputs. It further inc…
RAVEN is a LangGraph-powered copilot that helps equity analysts validate investment hypotheses fast. Paste a thesis and the agent assembles an end-to-end diligence loop: structured planning, data c…
A CLI tool for threat modeling and visualizing AI agents built using popular frameworks like LangGraph, AutoGen, CrewAI, and more.
Training Large Language Model to Reason in a Continuous Latent Space
aider is AI pair programming in your terminal
Lightweight coding agent that runs in your terminal
A free and strong UCI chess engine
A framework for few-shot evaluation of language models.
Performant framework for training, analyzing and visualizing Sparse Autoencoders (SAEs) and their frontier variants.
Code for NeurIPS 2024 paper "Regularizing Hidden States Enables Learning Generalizable Reward Model for LLMs"
A curated list of reinforcement learning with human feedback resources (continually updated)
Summaries for exciting works in the field of Deep Learning.