Tag

This paper introduces ReAct, an approach that integrates reasoning and acting in large language models (LLMs). ReAct enables LLMs to generate both reasoning traces and task-specific actions in an interleaved manner. This synergy allows reasoning to help induce, track, and update action plans, while actions interface with external sources like knowledge bases to gather more information, overcoming issues of hallucination and error propagation in prior methods.

SWE-agent is a system designed to empower language model (LM) agents to autonomously perform software engineering tasks. It features a custom agent-computer interface (ACI) that enhances the agent's ability to navigate repositories, create and edit code, and execute programs, achieving state-of-the-art results on the SWE-bench and HumanEvalFix benchmarks. [2, 5, 8]

Provides side-by-side, annotated PyTorch implementations of 60+ deep learning papers—transformers, diffusion models, GANs, optimizers, and RL algorithms. Runnable code pairs with explanatory notes on nn.labml.ai for learning and reproducing research.

A 57-subject multiple-choice benchmark for measuring broad language understanding in LLMs; provides per-subject configs and test/dev/auxiliary_train splits for few-/zero-shot evaluation, widely used for model comparison and academic reporting.

Benchmark dataset of ~8.5k grade-school math word problems with step-by-step solutions and calculator annotations for evaluating multi-step arithmetic reasoning in language models. Provided in two configs (main and socratic) and commonly used for chain-of-thought prompting, fine-tuning, and verifier training.

Provides human preference comparison pairs and red-team conversation transcripts collected by Anthropic for training preference/reward models and studying harmful model behaviors; intended for RLHF and safety research, not for supervised fine-tuning of dialogue agents.

Answers questions over scientific PDFs and other documents with citation-backed RAG; features metadata-aware embeddings, LLM re-ranking, multimodal PDF readers, and agentic workflows — suited for research QA and summarization but requires LLM/embedding access or local models.

Enables real-time (≥30 fps) 1080p novel-view synthesis by representing scenes as optimized anisotropic 3D Gaussians plus a visibility-aware splatting renderer; provides the paper's reference implementation, pretrained models and viewers — high-quality training requires CUDA GPU and significant VRAM.

Provides ~1.3 trillion tokens of web pages filtered for educational quality using an LLM-trained classifier; includes per-Crawl configs, smaller random samples (10B/100B/350B tokens), and the classifier code and model for reproducible filtering.

Teaches the foundations and recent advances of large language models through chapter-based PDFs and curated paper lists. Monthly updates keep the syllabus current; bilingual chapter materials and a structured paper list make it suitable for courses and self-study.

Enables agents to autonomously operate GUIs and complete complex computer tasks — includes the Agent S papers and the gui-agents SDK, grounding-model support, and runnable S3 agent implementations for Windows/macOS/Linux.

UI-TARS Desktop is a native desktop GUI agent by ByteDance that enables multimodal, vision-language-driven control of local and remote computers and browsers. It provides precise mouse/keyboard control, screenshot-based visual recognition, cross-platform support, and integration with the Agent TARS ecosystem and MCP tools. It focuses on private/local processing and building human-like task completion workflows.