📢🌟 Introducing SMART, a novel data mixture strategy for instruction tuning inspired by Submodularity! 🌟📢 📖 Pre-print available at arxiv.org/abs/2403.08370! 🤝 This work is in collaboration with Sumit Bhatia and Prof. Ganesh Ramakrishnan #ai #nlproc #instructiontuning #submodularity

📢🌟Thrilled to announce that our recent work SMART, a novel data mixture strategy for instruction tuning, has been accepted to ACL 2024(Findings)🌟📢 📖Pre-print at arxiv.org/abs/2403.08370! This work is in collaboration with Sumit Bhatia and Prof. Ganesh Ramakrishnan #nlproc #acl2024

We are hiring a Research Scientist in the field of Behavioral Sciences and Machine Learning at Adobe MDSR Labs. Requisites: PhD or equivalent experience Job Description: acrobat.adobe.com/id/urn:aaid:sc…

Sumit Bhatia is attending EMNLP 2025 Set up some time to talk to him to know more about the job and the lab #NLProc #EMNLP2024 #EMNLP

🚨 EMNLP 2024 🌴 Paper Alert! 🚨 aclanthology.org/2024.findings-… Introducing POSIX: A PrOmpt Sensitivity IndeX for Language Models! 🔎 Not all prompts are created equal...Small changes can lead to big surprises! - whether it's rephrasing, spelling errors, or even minor tweaks in

📢 Excited to share our latest research – “On the Effect of Instruction Tuning Loss on Generalization” – which will soon appear in the Transactions of the Association for Computational Linguistics (TACL)!! Instruction Tuning is the very first step in the post-pretraining phase