7 papers
Antidistillation Sampling
Yash Savani, Asher Trockman, Zhili Feng +5
Frontier models that generate extended reasoning traces inadvertently produce rich token sequences that can facilitate model distillation. Recognizing this vulnerability, model own…
Unnatural Languages Are Not Bugs but Features for LLMs
Keyu Duan, Yiran Zhao, Zhili Feng +9
Large Language Models (LLMs) have been observed to process non-human-readable text sequences, such as jailbreak prompts, often viewed as a bug for aligned LLMs. In this work, we pr…
Existing Large Language Model Unlearning Evaluations Are Inconclusive
Zhili Feng, Yixuan Even Xu, Alexander Robey +5
Machine unlearning aims to remove sensitive or undesired data from large language models. However, recent studies suggest that unlearning is often shallow, claiming that removed kn…
RankCLIP: Ranking-Consistent Language-Image Pretraining
Yiming Zhang, Zhuokai Zhao, Zhaorun Chen +3
Self-supervised contrastive learning models, such as CLIP, have set new benchmarks for vision-language models in many downstream tasks. However, their dependency on rigid one-to-on…
Rethinking LLM Memorization through the Lens of Adversarial Compression
Avi Schwarzschild, Zhili Feng, Pratyush Maini +2
Large language models (LLMs) trained on web-scale datasets raise substantial concerns regarding permissible data usage. One major question is whether these models "memorize" all th…
Adapting Language Models via Token Translation
Zhili Feng, Tanya Marwah, Nicolo Fusi +2
Modern large language models use a fixed tokenizer to effectively compress text drawn from a source domain. However, applying the same tokenizer to a new target domain often leads…