Zobrazeno 1 - 10
of 24
pro vyhledávání: '"Amini, Afra"'
Numerous previous studies have sought to determine to what extent language models, pretrained on natural language text, can serve as useful models of human cognition. In this paper, we are interested in the opposite question: whether we can directly
Externí odkaz:
http://arxiv.org/abs/2410.13086
Best-of-N (BoN) is a popular and effective algorithm for aligning language models to human preferences. The algorithm works as follows: at inference time, N samples are drawn from the language model, and the sample with the highest reward, as judged
Externí odkaz:
http://arxiv.org/abs/2407.06057
Autor:
Malagutti, Luca, Buinovskij, Andrius, Svete, Anej, Meister, Clara, Amini, Afra, Cotterell, Ryan
For nearly three decades, language models derived from the $n$-gram assumption held the state of the art on the task. The key to their success lay in the application of various smoothing techniques that served to combat overfitting. However, when neu
Externí odkaz:
http://arxiv.org/abs/2403.17240
Direct preference optimization (DPO) is a successful fine-tuning strategy for aligning large language models with human preferences without the need to train a reward model or employ reinforcement learning. DPO, as originally formulated, relies on bi
Externí odkaz:
http://arxiv.org/abs/2402.10571
Autor:
Du, Li, Amini, Afra, Hennigen, Lucas Torroba, Yu, Xinyan Velocity, Eisner, Jason, Lee, Holden, Cotterell, Ryan
Recent papers have demonstrated the possibility of energy-based text generation by adapting gradient-based sampling algorithms, a paradigm of MCMC algorithms that promises fast convergence. However, as we show in this paper, previous attempts on this
Externí odkaz:
http://arxiv.org/abs/2312.17710
Autor:
Bulian, Jannis, Schäfer, Mike S., Amini, Afra, Lam, Heidi, Ciaramita, Massimiliano, Gaiarin, Ben, Hübscher, Michelle Chen, Buck, Christian, Mede, Niels G., Leippold, Markus, Strauß, Nadine
Publikováno v:
Proceedings of the 41st International Conference on Machine Learning (ICML), 2024
As Large Language Models (LLMs) rise in popularity, it is necessary to assess their capability in critically relevant domains. We present a comprehensive evaluation framework, grounded in science communication research, to assess LLM responses to que
Externí odkaz:
http://arxiv.org/abs/2310.02932
We present a human-in-the-loop dashboard tailored to diagnosing potential spurious features that NLI models rely on for predictions. The dashboard enables users to generate diverse and challenging examples by drawing inspiration from GPT-3 suggestion
Externí odkaz:
http://arxiv.org/abs/2306.12146
We introduce a novel dependency parser, the hexatagger, that constructs dependency trees by tagging the words in a sentence with elements from a finite set of possible tags. In contrast to many approaches to dependency parsing, our approach is fully
Externí odkaz:
http://arxiv.org/abs/2306.05477
Gradient-based sampling algorithms have demonstrated their effectiveness in text generation, especially in the context of controlled text generation. However, there exists a lack of theoretically grounded and principled approaches for this task. In t
Externí odkaz:
http://arxiv.org/abs/2306.03061
Tasks that model the relation between pairs of tokens in a string are a vital part of understanding natural language. Such tasks, in general, require exhaustive pair-wise comparisons of tokens, thus having a quadratic runtime complexity in the length
Externí odkaz:
http://arxiv.org/abs/2305.15057