DecryptPrompt

Report Abuse

Basic Information

DecryptPrompt is a curated and continuously updated resource hub that collects and organizes literature, tutorials, and practical links around large language models, prompt engineering, and agent research. The repository groups papers, surveys, open-source models, datasets, frameworks, and blog posts into thematic sections such as chain-of-thought, RAG, RLHF, instruction tuning, memory systems, long-context handling, multimodal models, evaluation, inference optimization, and domain-specific LLMs. It serves as a reading guide and bibliography designed to help researchers, engineers, and advanced students navigate recent and classic work in LLM research and applied AIGC topics. The material is presented as categorized lists and topical writeups with pointers to implementation resources and explanatory blog series to facilitate study and project scoping.

Links

Categorization

App Details

Features
Extensive, well-organized bibliography and topical index covering many LLM subfields including prompt engineering, chain-of-thought, agents, RAG, RLHF, memory, long-input/output methods, multimodal models, quantization and MOE. Includes curated paper lists, surveys, reading-series blog posts, and links to open-source models, tool and agent frameworks, datasets, benchmarks and tutorials. The README is structured into many thematic sections for quick navigation and comparison of methods. It provides bilingual (primarily Chinese with references to English sources) commentary and practical pointers, and it is maintained as an evolving collection with frequent updates and additions to reflect new research and tools.
Use Cases
The repository helps practitioners and researchers perform literature reviews, discover key papers and surveys, and locate implementation resources such as models, datasets and frameworks relevant to specific LLM topics. It offers structured learning paths via themed sections and blog-series summaries that lower the barrier to understanding complex areas like agent tool use, retrieval-augmented generation, instruction fine-tuning and RLHF. Educators can use it as a syllabus or reading list. Engineers can use it to identify candidate open-source models, benchmarking datasets and engineering practices for building or evaluating LLM systems. The collection also aids in scoping research projects and finding pointers to practical code and experiments.

Please fill the required fields*