14 papers
Hidden in the Request: Explaining Unethical LLM Compliance through Token Relevance
Or Biton, Tomer Krichli, Itai Allouche +1
Although Large Language Models (LLMs) are aligned to optimize for both helpfulness and harmlessness, these dual objectives may conflict, inevitably leading to alignment failures. T…
Fully Differentiable Neural Forced Alignment via Soft Dynamic Programming
Rotem Rousso, Eyal Cohen, Joseph Keshet
Recent advances in sequence modeling have significantly improved ASR systems, bringing them close to human-level recognition accuracy and enhancing robustness across diverse acoust…
Why Sampling Is Not Choosing: Intentionality, Agency, and Moral Responsibility in Large Language Models
Joseph Keshet
Recent advances in large language models (LLMs) have prompted claims that such systems exhibit agency or qualify as moral agents. This paper argues that these attributions are misg…
Multilingual Word-Level Forced Alignment with Self-Supervised Representations and Learned Dynamic Programming
Roy Weber, Meidan Zehavi, Rotem Rousso +1
We present a method for accurate multilingual word-level forced alignment, consisting of an alignment encoder and a learned alignment decoder. The encoder integrates two representa…
Joint Enhancement and Classification using Coupled Diffusion Models of Signals and Logits
Gilad Nurko, Roi Benita, Yehoshua Dissen +4
Robust classification in noisy environments remains a fundamental challenge in machine learning. Standard approaches typically treat signal enhancement and classification as separa…
Analyzing and Guiding Zero-Shot Posterior Sampling in Diffusion Models
Roi Benita, Michael Elad, Joseph Keshet
Recovering a signal from its degraded measurements is a long standing challenge in science and engineering. Recently, zero-shot diffusion based methods have been proposed for such…