Efficient Test-Time Inference via Deterministic Exploration of Truncated Decoding Trees.

Johannes Zenn

Xueyan Li Johannes Zenn Ekaterina Fadeeva Guinan Su Mrinmaya Sachan Jonas Geiping Efficient Test-Time Inference via Deterministic Exploration of Truncated Decoding Trees. 2026 April abs/2604.20500 CoRR https://doi.org/10.48550/arXiv.2604.20500 db/journals/corr/corr2604.html#abs-2604-20500 streams/journals/corr

Johannes Zenn Jonas Geiping When are likely answers right? On Sequence Probability and Correctness in LLMs. 2026 June abs/2606.27359 CoRR https://doi.org/10.48550/arXiv.2606.27359 db/journals/corr/corr2606.html#abs-2606-27359 streams/journals/corr

Tim Z. Xiao Johannes Zenn Robert Bamler A Note on Generalization in Variational Autoencoders: How Effective Is Synthetic Data and Overparameterization? 2025 2025 Trans. Mach. Learn. Res. https://openreview.net/forum?id=bwyHf5eery db/journals/tmlr/tmlr2025.html#XiaoZB25

Tim Z. Xiao Johannes Zenn Zhen Liu 0019 Weiyang Liu Robert Bamler Bernhard Schölkopf Flipping Against All Odds: Reducing LLM Coin Flip Bias via Verbalized Rejection Sampling. 2025 June abs/2506.09998 CoRR https://doi.org/10.48550/arXiv.2506.09998 db/journals/corr/corr2506.html#abs-2506-09998 streams/journals/corr

Dillon Z. Chen Johannes Zenn Tristan Cinquin Sheila A. McIlraith Language Models For Generalised PDDL Planning: Synthesising Sound and Programmatic Policies. 2025 August abs/2508.18507 CoRR https://doi.org/10.48550/arXiv.2508.18507 db/journals/corr/corr2508.html#abs-2508-18507 streams/journals/corr

Johannes Zenn Robert Bamler Differentiable Annealed Importance Sampling Minimizes The Jensen-Shannon Divergence Between Initial and Target Distribution. 2024 ICML https://proceedings.mlr.press/v235/zenn24a.html https://openreview.net/forum?id=rvaN2P1rvC conf/icml/2024 db/conf/icml/icml2024.html#ZennB24 58448-58469

Johannes Zenn Robert Bamler Differentiable Annealed Importance Sampling Minimizes The Jensen-Shannon Divergence Between Initial and Target Distribution. 2024 abs/2405.14840 CoRR https://doi.org/10.48550/arXiv.2405.14840 db/journals/corr/corr2405.html#abs-2405-14840

Johannes Zenn Dominik Gond Fabian Jirasek Robert Bamler Balancing Molecular Information and Empirical Data in the Prediction of Physico-Chemical Properties. 2024 abs/2406.08075 CoRR https://doi.org/10.48550/arXiv.2406.08075 db/journals/corr/corr2406.html#abs-2406-08075

Johannes Zenn Robert Bamler Resampling Gradients Vanish in Differentiable Sequential Monte Carlo Samplers. 2023 Tiny Papers @ ICLR https://openreview.net/forum?id=kBkou5ucR_d conf/iclr/2023tiny db/conf/iclr/iclr2023tiny.html#ZennB23

Johannes Zenn Robert Bamler Resampling Gradients Vanish in Differentiable Sequential Monte Carlo Samplers. 2023 abs/2304.14390 CoRR https://doi.org/10.48550/arXiv.2304.14390 db/journals/corr/corr2304.html#abs-2304-14390

Tim Z. Xiao Johannes Zenn Robert Bamler Upgrading VAE Training With Unlimited Data Plans Provided by Diffusion Models. 2023 abs/2310.19653 CoRR https://doi.org/10.48550/arXiv.2310.19653 db/journals/corr/corr2310.html#abs-2310-19653

Tim Z. Xiao Johannes Zenn Robert Bamler The SVHN Dataset Is Deceptive for Probabilistic Generative Models Due to a Distribution Mismatch. 2023 abs/2312.02168 CoRR https://doi.org/10.48550/arXiv.2312.02168 db/journals/corr/corr2312.html#abs-2312-02168

Jonathan Wenger Nicholas Krämer Marvin Pförtner Jonathan Schmidt 0001 Nathanael Bosch Nina Effenberger Johannes Zenn Alexandra Gessner Toni Karvonen François-Xavier Briol Maren Mahsereci Philipp Hennig ProbNum: Probabilistic Numerics in Python. 2021 abs/2112.02100 CoRR https://arxiv.org/abs/2112.02100 db/journals/corr/corr2112.html#abs-2112-02100

Robert Bamler Nathanael Bosch François-Xavier Briol Dillon Ze ChenDillon Z. Chen Tristan Cinquin Nina Effenberger Ekaterina Fadeeva Jonas Geiping Alexandra Gessner Dominik Gond Philipp Hennig Fabian Jirasek Toni Karvonen Nicholas Krämer Xueyan Li Weiyang Liu Zhen Liu 0019 Maren Mahsereci Sheila A. McIlraith Marvin Pförtner Mrinmaya Sachan Jonathan Schmidt 0001 Bernhard Schölkopf Guinan Su Jonathan Wenger Tim Z. Xiao