Jonas Geiping

MedSAMix: A Training-Free Model Merging Approach for Medical Image Segmentation AAAI 2026 An Interpretable N-gram Perplexity Threat Model for Large Language Model Jailbreaks ICML 2025 Great Models Think Alike and this Undermines AI Oversight ICML 2025 LLM-Generated Passphrases That Are Secure and Easy to Remember NAACL 2025 When, Where and Why to Average Weights? ICML 2025 Be like a Goldfish, Don't Memorize! Mitigating Memorization in Generative LLMs NeurIPS 2024 CALVIN: Improved Contextual Video Captioning via Instruction Tuning NeurIPS 2024 NEFTune: Noisy Embeddings Improve Instruction Finetuning ICLR 2024 Object Recognition as Next Token Prediction CVPR 2024 On the Reliability of Watermarks for Large Language Models ICLR 2024 Privacy Backdoors: Enhancing Membership Inference through Poisoning Pre-trained Models NeurIPS 2024 Spotting LLMs With Binoculars: Zero-Shot Detection of Machine-Generated Text ICML 2024 Transformers Can Do Arithmetic with the Right Embeddings NeurIPS 2024 Universal Guidance for Diffusion Models ICLR 2024 A Performance-Driven Benchmark for Feature Selection in Tabular Deep Learning NeurIPS 2023 Augmenters at SemEval-2023 Task 1: Enhancing CLIP in Handling Compositionality and Ambiguity for Zero-Shot Visual WSD through Prompt Augmentation and Text-To-Image Diffusion SemEval 2023 A Watermark for Large Language Models ICML 2023 Canary in a Coalmine: Better Membership Inference with Ensembled Adversarial Queries ICLR 2023 Cold Diffusion: Inverting Arbitrary Image Transforms Without Noise NeurIPS 2023 Cramming: Training a Language Model on a single GPU in one day. ICML 2023 Decepticons: Corrupted Transformers Breach Privacy in Federated Learning for Language Models ICLR 2023 Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion Models CVPR 2023 Hard Prompts Made Easy: Gradient-Based Discrete Optimization for Prompt Tuning and Discovery NeurIPS 2023 How Much Data Are Augmentations Worth? An Investigation into Scaling Laws, Invariance, and Implicit Regularization ICLR 2023 Loss Landscapes are All You Need: Neural Network Generalization Can Be Explained Without the Implicit Bias of Gradient Descent ICLR 2023 On the Exploitability of Instruction Tuning NeurIPS 2023 Panning for Gold in Federated Learning: Targeted Text Extraction under Arbitrarily Large-Scale Aggregation ICLR 2023 Tree-Rings Watermarks: Invisible Fingerprints for Diffusion Images NeurIPS 2023 Understanding and Mitigating Copying in Diffusion Models NeurIPS 2023 What Can We Learn from Unlearnable Datasets? NeurIPS 2023 Autoregressive Perturbations for Data Poisoning NeurIPS 2022 Robbing the Fed: Directly Obtaining Private Data in Federated Learning with Modified Models ICLR 2022 Stochastic Training is Not Necessary for Generalization ICLR 2022 Adversarial Examples Make Strong Poisons NeurIPS 2021 Witches' Brew: Industrial Scale Data Poisoning via Gradient Matching ICLR 2021 Inverting Gradients - How easy is it to break privacy in federated learning? NeurIPS 2020 MetaPoison: Practical General-purpose Clean-label Data Poisoning NeurIPS 2020 Truth or backpropaganda? An empirical investigation of deep learning theory ICLR 2020 Parametric Majorization for Data-Driven Energy Minimization Methods ICCV 2019