Testing theory of mind in large language models and humans - Nature Human Behaviour
nature.com
Testing theory of mind in large language models and humans - Nature Human Behaviour
Maybe human intelligence is more illusory than we think, a kind of egotistical fantasy. Are human authors really very different from those Balnibarbian pupils spinning the cranks of their writing machine? Human authors learn to write by reading other books in their chosen genre (“training”), and maybe what’s happening in their brains is just a slig
... See moreOrion Magazine • The Magic in the Machine
Observing these patterns of influence gives clues about how our models generalize from their training data. For instance, if the models responded to user prompts by splicing together sequences from the training set, then we’d expect the influential sequences for a given model response to include expressions of near-identical thoughts. Conversely, i... See more
Anthropic \ Tracing Model Outputs to the Training Data
The community remains puzzled about whether these models genuinely generalize to unseen tasks, or seemingly succeed by memorizing the training data. This paper makes important strides in addressing this question. It constructs a suite of carefully designed counterfactual evaluations, providing fresh insights into the capabilities of state-of-the-ar... See more
Zhaofeng Wu • Reasoning skills of large language models are often overestimated
In the work with LLMs lies a deeper question, that I have not been able to answer for myself properly so far. I can use an LLM to extract information, summarise, distill, but it feels like I am fooling myself. The work that I have done is not done by me. I am have not made any connections in my brain. It feels like we should not use them for intell
... See more