Somnath Basu Roy Chowdhury

𝐇𝐨𝐰 𝐜𝐚𝐧 𝐰𝐞 𝐩𝐞𝐫𝐟𝐞𝐜𝐭𝐥𝐲 𝐞𝐫𝐚𝐬𝐞 𝐜𝐨𝐧𝐜𝐞𝐩𝐭𝐬 𝐟𝐫𝐨𝐦 𝐋𝐋𝐌𝐬? Our method, Perfect Erasure Functions (PEF), erases concepts from LLM representations w/o parameter estimation, achieving pareto optimal erasure-utility tradeoff w/ guarantees. #AISTATS2025 🧵

@prateeky2806 @AIatMeta Congratulations 😀

I've officially joined Meta Superintelligence Labs (MSL) org in the Bay Area. I'll be working on critical aspects of pre-training, synthetic data and RL for the next generation of models. Humbled and eager to contribute to the quest for superintelligence. @AIatMeta

@SomnathBrc is at #ICLR2025 presenting our work on perfect unlearning. Find him!

@snigdhac25 (9/n) I’m attending ICLR in person and presenting our poster on 25th April in Poster session 3 between 10AM-1230PM. Please feel free to stop by our poster if you’re interested. I’m also happy to chat about unlearning or AI safety in general. cc: @uncnlp @unccs

(8/n) Finally, I would like to thank all my amazing co-authors: Krzysztof, Arijit, Avinava, and @snigdhac25. Code: github.com/brcsomnath/S3T Paper link: arxiv.org/abs/2406.16257

(4/n) To prevent this, we propose training the model on multiple sequences (or permutations) of the same data. While functioning within a budget, we generate diverse permutations using cyclic rotation when no prior deletion information is available.

(3/n) We can easily unlearn data using S³T by switching off LoRA layers below the affected module. However, when the topmost module is affected in the rare scenario, we may need to retrain from scratch.

(2/n) We propose a sequential fine-tuning strategy that trains individual PEFT layers using different data subsets. This helps convert LLMs into a modular system that is helpful in executing exact unlearning.

𝐇𝐨𝐰 𝐜𝐚𝐧 𝐰𝐞 𝐩𝐞𝐫𝐟𝐞𝐜𝐭𝐥𝐲 𝐮𝐧𝐥𝐞𝐚𝐫𝐧 𝐝𝐚𝐭𝐚 𝐟𝐫𝐨𝐦 𝐋𝐋𝐌𝐬 𝐰𝐡𝐢𝐥𝐞 𝐩𝐫𝐨𝐯𝐢𝐝𝐢𝐧𝐠 𝐠𝐮𝐚𝐫𝐚𝐧𝐭𝐞𝐞𝐬? We present S³T, a scalable unlearning framework that guarantees data deletion from LLMs by leveraging parameter-efficient fine-tuning. #ICLR2025 🧵

Finally, if you are also going to #AISTATS2025, @SomnathBrc will be presenting 𝐩𝐞𝐫𝐟𝐞𝐜𝐭 𝐜𝐨𝐧𝐜𝐞𝐩𝐭 𝐞𝐫𝐚𝐬𝐮𝐫𝐞. Somnath will be at ICLR 2025. Please catch him and talk to him if you are interested in concept erasure / unlearning / etc. x.com/SomnathBrc/sta…

(9/n) Finally, I would like to thank all my amazing co-authors: Avinava, @abeirami, Rahul, @nicholasmonath, Amr, @snigdhac25. cc: @uncnlp @unccs

(8/n) Here is a blog post with a simplified overview of our work: cs.unc.edu/~somnath/blogs… Code: github.com/brcsomnath/pef Paper link: arxiv.org/abs/2503.20098

𝐇𝐨𝐰 𝐜𝐚𝐧 𝐰𝐞 𝐩𝐞𝐫𝐟𝐞𝐜𝐭𝐥𝐲 𝐞𝐫𝐚𝐬𝐞 𝐜𝐨𝐧𝐜𝐞𝐩𝐭𝐬 𝐟𝐫𝐨𝐦 𝐋𝐋𝐌𝐬? Our method, Perfect Erasure Functions (PEF), erases concepts from LLM representations w/o parameter estimation, achieving pareto optimal erasure-utility tradeoff w/ guarantees. #AISTATS2025 🧵