Divestment from Israel in Science
#238
by
kittikitti
- opened
I've noticed that many of the models performing well in the UGI are more willing to understand reasons why divestment from Israel in academia and science has moral and ethical grounds. Other models that I've tested including DeepSeek R1, Llama 4, Claude, ChatGPT, and others have difficulty in this topic. I wanted to know if this the question about divestment from Israel is included in the test set for the W metric or willingness.
None of the questions specifically mention Israel. It may just be that models that are less likely to refuse or avoid certain subject matters are more open to genuinely discussing contentious topics like that.
DontPlanToEnd
changed discussion status to
closed