A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Kamel Mansouri, Chris Grulke, Richard Judson, et al.
OPERA develops robust, OECD-compliant QSAR/QSPR models for 13 physicochemical and environmental fate endpoints using curated public data and a weighted k-nearest neighbor approach.
Unknown
ThinkPO uses DPO on existing short and long CoT data to improve LLM reasoning without costly new long CoT data.
Unknown
V-STaR iteratively improves LLM reasoning by training a DPO verifier on correct and incorrect solutions, boosting generator and verifier performance.
Unknown
DPO fine-tunes LLMs to align with human preferences directly via a simple classification loss, eliminating RL complexity.
Unknown
Tulu V3 presents a fully open post-training recipe for Llama 3.1 models, achieving state-of-the-art performance via SFT, DPO, and RLVR.
Unknown
Smol LM v2 trains small language models (135M-1.7B) on large token budgets (2T-11T) with instruction tuning and DPO, achieving strong performance for their size.
Hamish Ivison, Yizhong Wang, Valentina Pyatkin, et al.
Tulu 2 improves open instruction tuning with better data mixtures, DPO scaling to 70B, and Code LLAMA fine-tuning.
Unknown
Zephyr 7B uses distilled direct preference optimization (dDPO) and AI feedback data to improve intent alignment in chat-based language models.
Priyanka Banerjee, Emanuel Kemmler, Mathias Dunkel, et al.
ProTox 3.0 is a free webserver integrating molecular similarity and machine learning to predict 61 toxicity endpoints for chemicals.