[Roadmap] Interesting research papers integrate

[lightaime]

Required prerequisites

• I have searched the Issue Tracker and Discussions that this hasn't already been reported. (+1 or comment there if it has.)
• Consider asking first in a Discussion.

Motivation

Data generation:

• ReBase: https://x.com/aomaru_21490/status/1810553283335143711?s=46
• APIGen: https://arxiv.org/abs/2406.18518
• Chain-of-Instructions: https://arxiv.org/pdf/2402.11532
• Distilling System 2 into System 1: https://arxiv.org/pdf/2407.06023v2
• Dart-math: https://github.com/hkust-nlp/dart-math
• Rephrasing the Web: https://x.com/pratyushmaini/status/1752337225097076809?s=46
• DataComp-LM: https://arxiv.org/abs/2406.11794
• Self-play with Execution Feedback: https://x.com/keminglu612/status/1803996449644122316?s=46
• STaR: https://arxiv.org/abs/2203.14465
• DeepSeek-Prover: https://arxiv.org/pdf/2405.14333
• Magpie: https://arxiv.org/abs/2406.08464
• RAGEval: https://x.com/omarsar0/status/1820507831491239978?s=46
• MINT-1T: https://x.com/arankomatsuzaki/status/1802907035236704667?s=46
• Best Practices and Lessons Learned on Synthetic Data for Language Models: https://arxiv.org/pdf/2404.07503v1
• More: https://huggingface.co/collections/stereoplegic/dataset-generation-65389dd75510eb595f8a3797

Data curation:

• TracIn: https://arxiv.org/abs/2405.17490
• CKNN-Shapley: https://arxiv.org/pdf/2405.17489
• More: https://github.com/tongyx361/Awesome-LLM4Math

Multi-modality:

• Anole: https://x.com/stefan_fee/status/1810695036432232576?s=46
• InternLM-XComposer-2.5: https://x.com/wjqdev/status/1808751706970489287?s=46
• Mira: https://mira-space.github.io/
• More: Efficient Multimodal Large Language Models:A Survey: https://github.com/lijiannuist/Efficient-Multimodal-LLMs-Survey

RAG:

• GraphEval: https://x.com/hitesh_lpatel/status/1813335989034635515?s=46
• G-Retriever: https://arxiv.org/pdf/2402.07630

Reward model / judge:

• MJ-Bench: https://x.com/huaxiuyaoml/status/1810728309182861462?s=46
• Synthetic Critiques: https://arxiv.org/pdf/2405.20850

Agent:

• EvoAgent: https://arxiv.org/abs/2406.14228
• CodeNav: https://x.com/tanmay2099/status/1806020668175376809?s=46
• AgentGym: https://arxiv.org/abs/2406.04151
• LLM Reasoners: https://x.com/maitrixorg/status/1782801835423850881?s=46
• Structured output: https://simmering.dev/blog/structured_output/

Efficiency:

• SGLang: https://arxiv.org/abs/2312.07104

Prompt Optimization

• DSPy: https://github.com/stanfordnlp/dspy
• TextGrad: https://github.com/zou-group/textgrad

Benchmark:

• ARC: https://www.kaggle.com/competitions/arc-prize-2024/
• StableToolBench: https://github.com/THUNLP-MT/StableToolBench
• CodeRAG-Bench: https://x.com/zhiruow/status/1804138112303378487?s=46
• Berkeley Function-Calling Leaderboard: https://gorilla.cs.berkeley.edu/leaderboard.html#leaderboard
• R2E: Repository to Environment: https://x.com/slimshetty_/status/1787860906439029022?s=46

Solution

No response

Alternatives

No response

Additional context

No response