| Introducing MPT-7B: A new standard for open-source, commercially usable LLMs MosaicML NLP Team https://www.databricks.com/blog/mpt-7b, 05-05, 2023 | 283* | 2023 |
| Introducing MPT-30B: Raising the bar for open-source foundation models MosaicML NLP Team Accessed, 2023 | 45 | 2023 |
| Practical leverage-based sampling for low-rank tensor decomposition BW Larsen, TG Kolda SIAM Journal on Matrix Analysis and Applications 43 (3), 1488-1517, 2022 | 45 | 2022 |
| Unmasking the Lottery Ticket Hypothesis: What's Encoded in a Winning Ticket's Mask? M Paul*, F Chen*, BW Larsen*, J Frankle, S Ganguli, GK Dziugaite International Conference on Learning Representations 2023, 2022 | 44 | 2022 |
| How many degrees of freedom do we need to train deep networks: a loss landscape perspective BW Larsen, S Fort, N Becker, S Ganguli International Conference on Learning Representations 2022, 2021 | 24 | 2021 |
| Introducing DBRX: A new state-of-the-art open LLM Mosaic Research Team URL https://www. databricks. com/blog/introducing-dbrx-new-state-art-open …, 2024 | 20 | 2024 |
| Lottery tickets on a data diet: Finding initializations with sparse trainable networks M Paul*, B Larsen*, S Ganguli, J Frankle, GK Dziugaite Advances in Neural Information Processing Systems 35, 18916-18928, 2022 | 19 | 2022 |
| Duality of Bures and Shape Distances with Implications for Comparing Neural Representations SE Harvey, BW Larsen, AH Williams Proceedings of UniReps: the First Workshop on Unifying Representations in …, 2024 | 13 | 2024 |
| Avoiding spurious local minima in deep quadratic networks A Kazemipour, BW Larsen, S Druckmann arXiv preprint arXiv:2001.00098, 2019 | 12* | 2019 |
| Does your data spark joy? Performance gains from domain upsampling at the end of training C Blakeney*, M Paul*, BW Larsen*, S Owen, J Frankle Conference on Language Models 2024, arXiv: 2406.03476, 2024 | 11* | 2024 |
| Estimating shape distances on neural representations with limited samples DA Pospisil, BW Larsen, SE Harvey, AH Williams International Conference on Learning Representations 2024, 2023 | 5 | 2023 |
| Applying matching pursuit decomposition time-frequency processing to UGS footstep classification BW Larsen, H Chung, A Dominguez, J Sciacca, N Kovvali, ... Sensors, and Command, Control, Communications, and Intelligence (C3I …, 2013 | 4 | 2013 |
| Towards a more general understanding of the algorithmic utility of recurrent connections BW Larsen, S Druckmann PLOS Computational Biology 18 (6), e1010227, 2022 | 2 | 2022 |
| Sketching matrix least squares via leverage scores estimates BW Larsen, TG Kolda arXiv preprint arXiv:2201.10638, 2022 | 2 | 2022 |
| Unmasking the Lottery Ticket Hypothesis: Efficient Adaptive Pruning for Finding Winning Tickets M Paul, F Chen, BW Larsen, J Frankle, S Ganguli, GK Dziugaite Has it Trained Yet? NeurIPS 2022 Workshop, 2022 | 2 | 2022 |
| Tensor decomposition meets rkhs: Efficient algorithms for smooth and misaligned data BW Larsen, TG Kolda, AR Zhang, AH Williams arXiv preprint arXiv:2408.05677, 2024 | 1 | 2024 |
| Pre-Training on a Data Diet: Identifying Sufficient Examples for Early Training M Paul, BW Larsen, S Ganguli, J Frankle, GK Dziugaite First Workshop on Pre-training: Perspectives, Pitfalls, and Paths Forward at …, 2022 | 1 | 2022 |
| Three-dimensional imaging utilizing low frequency magnetic fields DR Allee, GP Spell, B Larsen, AM Wilson, OC Ma US Patent 10,416,244, 2019 | 1 | 2019 |
| Optimization and High-Dimensional Loss Landscapes in Deep Learning BW Larsen Stanford University, 2022 | | 2022 |
| Developing a Flexible Electric and Magnetic Field Imaging Blanket BW Larsen | | 2015 |
Surya GanguliAssociate Professor, Stanford UniversityEmail verificata su stanford.edu