Linearly mapping from image to text space J Merullo, L Castricato, C Eickhoff, E Pavlick arXiv preprint arXiv:2209.15162, 2022 | 64 | 2022 |
Language models implement simple word2vec-style vector arithmetic J Merullo, C Eickhoff, E Pavlick arXiv preprint arXiv:2305.16130, 2023 | 23* | 2023 |
Investigating sports commentator bias within a large corpus of American football broadcasts J Merullo, L Yeh, A Handler, A Grissom II, B O'Connor, M Iyyer arXiv preprint arXiv:1909.03343, 2019 | 17 | 2019 |
Does clip bind concepts? probing compositionality in large image models M Lewis, NV Nayak, P Yu, Q Yu, J Merullo, SH Bach, E Pavlick arXiv preprint arXiv:2212.10537, 2022 | 16 | 2022 |
Characterizing mechanisms for factual recall in language models Q Yu, J Merullo, E Pavlick arXiv preprint arXiv:2310.15910, 2023 | 11 | 2023 |
Circuit component reuse across tasks in transformer language models J Merullo, C Eickhoff, E Pavlick arXiv preprint arXiv:2310.08744, 2023 | 9 | 2023 |
ezCoref: Towards unifying annotation guidelines for coreference resolution A Gupta, M Karpinska, W Zhao, K Krishna, J Merullo, L Yeh, M Iyyer, ... arXiv preprint arXiv:2210.07188, 2022 | 3 | 2022 |
Pretraining on interactions for learning grounded affordance representations J Merullo, D Ebert, C Eickhoff, E Pavlick arXiv preprint arXiv:2207.02272, 2022 | 3 | 2022 |
Transformer Mechanisms Mimic Frontostriatal Gating Operations When Trained on Human Working Memory Tasks A Traylor, J Merullo, MJ Frank, E Pavlick arXiv preprint arXiv:2402.08211, 2024 | 1 | 2024 |
ACQuA: Arrhythmia Classification with Quasi-Attractors W Rudman, J Merullo, L Mercurio, C Eickhoff medRxiv, 2022.08. 31.22279436, 2022 | 1* | 2022 |
Axiomatic Causal Interventions for Reverse Engineering Relevance Computation in Neural Retrieval Models C Chen, J Merullo, C Eickhoff arXiv preprint arXiv:2405.02503, 2024 | | 2024 |