Zobrazeno 1 - 7
of 7
pro vyhledávání: '"Ruis, Laura"'
Autor:
Ruis, Laura, Mozes, Maximilian, Bae, Juhan, Kamalakara, Siddhartha Rao, Talupuru, Dwarak, Locatelli, Acyr, Kirk, Robert, Rocktäschel, Tim, Grefenstette, Edward, Bartolo, Max
The capabilities and limitations of Large Language Models have been sketched out in great detail in recent years, providing an intriguing yet conflicting picture. On the one hand, LLMs demonstrate a general ability to solve problems. On the other han
Externí odkaz:
http://arxiv.org/abs/2411.12580
Representation engineering methods have recently shown promise for enabling efficient steering of model behavior. However, evaluation pipelines for these methods have primarily relied on subjective demonstrations, instead of quantitative, objective m
Externí odkaz:
http://arxiv.org/abs/2410.17245
Autor:
Khan, Akbir, Hughes, John, Valentine, Dan, Ruis, Laura, Sachan, Kshitij, Radhakrishnan, Ansh, Grefenstette, Edward, Bowman, Samuel R., Rocktäschel, Tim, Perez, Ethan
Common methods for aligning large language models (LLMs) with desired behaviour heavily rely on human-labelled data. However, as models grow increasingly sophisticated, they will surpass human expertise, and the role of human evaluation will evolve i
Externí odkaz:
http://arxiv.org/abs/2402.06782
Autor:
Ruis, Laura, Khan, Akbir, Biderman, Stella, Hooker, Sara, Rocktäschel, Tim, Grefenstette, Edward
Despite widespread use of LLMs as conversational agents, evaluations of performance fail to capture a crucial aspect of communication: interpreting language in context -- incorporating its pragmatics. Humans interpret language using beliefs and prior
Externí odkaz:
http://arxiv.org/abs/2210.14986
Autor:
Ruis, Laura, Lake, Brenden
Systematic generalization is the ability to combine known parts into novel meaning; an important aspect of efficient human learning, but a weakness of neural network learning. In this work, we investigate how two well-known modeling principles -- mod
Externí odkaz:
http://arxiv.org/abs/2202.10745
Humans easily interpret expressions that describe unfamiliar situations composed from familiar parts ("greet the pink brontosaurus by the ferris wheel"). Modern neural networks, by contrast, struggle to interpret novel compositions. In this paper, we
Externí odkaz:
http://arxiv.org/abs/2003.05161
We propose the Insertion-Deletion Transformer, a novel transformer-based neural architecture and training method for sequence generation. The model consists of two phases that are executed iteratively, 1) an insertion phase and 2) a deletion phase. T
Externí odkaz:
http://arxiv.org/abs/2001.05540