Publications of T Saanum

Binz, M.; Akata, E.; Bethge, M.; Brändle, F.; Callaway, F.; Coda-Forno, J.; Dayan, P.; Demircan, C.; Eckstein, M.; Éltetö, N. et al.; Griffiths, T.; Haridi, S.; Jagadish, A.; Li, J.-A.; Kipnis, A.; Kumar, S.; Ludwig, T.; Mathony, M.; Mattar, M.; Modirshanechi, A.; Nath, S.; Peterson, J.; Rmus, M.; Russek, E.; Saanum, T.; Schubert, J.; Schulze Buschoff, L.; Singhi, N.; Sui, X.; Thalmann, M.; Theis, F.; Truong, V.; Udandarao, V.; Voudouris, K.; Wilson, R.; Witte, K.; Wu, S.; Wulff, D.; Xiong, H.; Schulz, E.: A foundation model to predict and capture human cognition. Nature 644 (8078), pp. 1002 - 1009 (2025)

Garvert, M.; Saanum, T.; Schulz, E.; Schuck, N.; Doeller, C.: Hippocampal spatio-predictive cognitive maps adaptively guide reward generalization. Nature Neuroscience 26 (4), pp. 615 - 626 (2023)

Demircan, C.; Saanum, T.; Pettini, L.; Binz, M.; Baczkowski, B.; Doeller, C.; Garvert, M.; Schulz, E.: Evaluating alignment between humans and neural network representations in image-based learning tasks. In: NIPS '24: Proceedings of the 38th International Conference on Neural Information Processing Systems, 3890, pp. 122406 - 122433 (Eds. Globerson, A.; Mackey, L.; Belgrave, D.; Fan, A.; Paquet, U. et al.). Thirty-Eighth Annual Conference on Neural Information Processing Systems (NeurIPS 2024), Vancouver, Canada, December 11, 2024 - December 15, 2024. Curran, Red Hook, NY, USA (2025)

Saanum, T.; Dayan, P.; Schulz, E.: Simplifying Latent Dynamics with Softly State-Invariant World Models. In: NIPS '24: Proceedings of the 38th International Conference on Neural Information Processing Systems, 1212, pp. 38355 - 38382 (Eds. Globerson, A.; Mackay, L.; Belgrave, D.; Fan, A.; Paquet, U. et al.). Thirty-eighth Conference on Neural Information Processing Systems (NeurIPS 2024), Vancouver, Canada, December 10, 2024 - December 15, 2024. Curran, Red Hook, NY, USA (2025)

Demircan, C.; Saanum, T.; Jagadish, A.; Binz, M.; Schulz, E.: Sparse Autoencoders Reveal Temporal Difference Learning in Large Language Models. In: Thirteenth International Conference on Learning Representations (ICLR 2025). Thirteenth International Conference on Learning Representations (ICLR 2025), Singapore, April 24, 2025 - April 28, 2025. (2025)

Saanum, T.; Schulze Buschoff, L.; Dayan, P.; Schulz, E.: Next state prediction gives rise to entangled, yet compositional representations of objects. In: Thirteenth International Conference on Learning Representations (ICLR 2025). Thirteenth International Conference on Learning Representations (ICLR 2025), Singapore, April 24, 2025 - April 28, 2025. (2025)

Saanum, T.; Éltetö, N.; Dayan, P.; Binz, M.; Schulz, E.: Reinforcement Learning with Simple Sequence Priors. In: Advances in Neural Information Processing Systems 36: 37th Conference on Neural Information Processing Systems (NeurIPS 2023), 2710, pp. 61985 - 62005 (Eds. Oh, A.; Naumann, T.; Globerson, A.; Saenko, K.; Hardt, M. et al.). Thirty-seventh Conference on Neural Information Processing Systems (NeurIPS 2023), New Orleans, LA, USA, December 10, 2023 - December 16, 2023. Curran, Red Hook, NY, USA (2024)

Demircan, C.; Pettini, L.; Saanum, T.; Binz, M.; Baczkowski, B.; Doeller, C.; Garvert, M.; Schulz, E.: Decision-Making with Naturalistic Options. In: 44th Annual Meeting of the Cognitive Science Society (CogSci 2022): Cognitive Diversity, pp. 976 - 982 (Eds. Culbertson, J.; Perfors, A.; Rabagliati, H.; Ramenzoni, V.). 44th Annual Meeting of the Cognitive Science Society (CogSci 2022): Cognitive Diversity, Toronto, Canada, July 27, 2022 - July 30, 2022. (2022)

Jagadish, A.; Saanum, T.; Wang, J.; Binz, M.; Schulz, E.: Probing Compositional Inference in Natural and Artificial Agents. In: 5th Multidisciplinary Conference on Reinforcement Learning and Decision Making (RLDM 2022), 1.67, pp. 275 - 279. 5th Multidisciplinary Conference on Reinforcement Learning and Decision Making (RLDM 2022), Providence, RI, USA, June 08, 2022 - June 11, 2022. (2022)

Saanum, T.; Schulz, E.; Speekenbrink, M.: Compositional generalization in multi-armed bandits. In: 43rd Annual Conference of the Cognitive Science Society (CogSci 2021): Workshop 2 Using Games to Understand Intelligence, pp. 1320 - 1326. 43rd Annual Conference of the Cognitive Science Society (CogSci 2021), Wien, Austria, July 26, 2021 - July 29, 2021. Curran, Red Hook, NY, USA (2021)

Saanum, T.; Schulz, E.: Not all induction heads are created equal. 8th Annual Conference on Cognitive Computational Neuroscience (CCN 2025), Amsterdam, The Netherlands (2025)

Saanum, T.; Schulz, E.: Learning parsimonious dynamics for state abstraction and navigation. Computational and Systems Neuroscience Meeting (COSYNE 2023), Montreal, Quebec, Canada (2023)

Saanum, T.; Garvert, M.; Schulz, E.; Schuck, N.; Doeller, C.: Hippocampal spatio-temporal cognitive maps adaptively guide reward generalization. Computational and Systems Neuroscience Meeting (COSYNE 2022), Lisboa, Portugal (2022)

Binz, M.; Akata, E.; Bethge, M.; Brändle, F.; Callaway, F.; Coda-Forno, J.; Dayan, P.; Demircan, C.; Eckstein, M.; Éltetö, N. et al.; Griffiths, T.; Haridi, S.; Jagadish, A.; Ji-An; Kipnis, A.; Kumar, S.; Ludwig, T.; Mathony, M.; Mattar, M.; Modirshanechi, A.; Nath, S.; Peterson, J.; Rmus, M.; Russek, E.; Saanum, T.; Scharfenberg, N.; Schubert, J.; Schulze Buschoff, L.; Singhi, N.; Sui, X.; Thalmann, M.; Theis, F.; Truong, V.; Udandarao, V.; Voudouris, K.; Wilson, R.; Witte, K.; Wu, S.; Wulff, D.; Xiong, H.; Schulz, E.: Centaur: a foundation model of human cognition. (submitted)

Demircan, C.; Saanum, T.; Jagadish, A.; Binz, M.; Schulz, E.: Sparse Autoencoders Reveal Temporal Difference Learning in Large Language Models. (submitted)

Jagadish, A.; Binz, M.; Saanum, T.; Wang, J.; Schulz, E.: Zero-shot compositional reinforcement learning in humans. (submitted)

Demircan, C.; Saanum, T.; Pettini, L.; Binz, M.; Baczkowski, B.; Kaanders, P.; Doeller, C.; Garvert, M.; Schulz, E.: Language Aligned Visual Representations Predict Human Behavior in Naturalistic Learning Tasks. (submitted)

Saanum, T.; Schulz, E.: Learning Parsimonious Dynamics for Generalization in Reinforcement Learning. (submitted)