Summary of Datadreamer: a Tool For Synthetic Data Generation and Reproducible Llm Workflows, by Ajay Patel et al.
DataDreamer: A Tool for Synthetic Data Generation and Reproducible LLM Workflowsby Ajay Patel, Colin Raffel,…
DataDreamer: A Tool for Synthetic Data Generation and Reproducible LLM Workflowsby Ajay Patel, Colin Raffel,…
Subgraph-level Universal Prompt Tuningby Junhyun Lee, Wooseong Yang, Jaewoo KangFirst submitted to arxiv on: 16…
BitDelta: Your Fine-Tune May Only Be Worth One Bitby James Liu, Guangxuan Xiao, Kai Li,…
Recovering the Pre-Fine-Tuning Weights of Generative Modelsby Eliahu Horwitz, Jonathan Kahana, Yedid HoshenFirst submitted to…
Rewards-in-Context: Multi-objective Alignment of Foundation Models with Dynamic Preference Adjustmentby Rui Yang, Xiaoman Pan, Feng…
Self-Play Fine-Tuning of Diffusion Models for Text-to-Image Generationby Huizhuo Yuan, Zixiang Chen, Kaixuan Ji, Quanquan…
A StrongREJECT for Empty Jailbreaksby Alexandra Souly, Qingyuan Lu, Dillon Bowen, Tu Trinh, Elvis Hsieh,…
MIM-Refiner: A Contrastive Learning Boost from Intermediate Pre-Trained Representationsby Benedikt Alkin, Lukas Miklautz, Sepp Hochreiter,…
Parameter-tuning-free data entry error unlearning with adaptive selective synaptic dampeningby Stefan Schoepf, Jack Foster, Alexandra…
LoraRetriever: Input-Aware LoRA Retrieval and Composition for Mixed Tasks in the Wildby Ziyu Zhao, Leilei…