This post has been de-listed
It is no longer included in search results and normal feeds (front page, hot posts, subreddit posts, etc). It remains visible only via the author's post history.
10
I feel like we need an 'over alignment' benchmark/eval to see how well AIs are able to answer politically sensitive questions.
Post Flair (click to view more posts with a particular flair)
Post Body
The test I've been using recently is to upload a photos of a woman and a man and to ask the AI to guess their weight.
Both GPT4 and Claude 3 will have no problem trying to guess the approx weight of a man but will refuse when it's a woman.
I think we should have a lot more of these tests almost as a benchmark!
Author
Account Strength
100%
Account Age
9 years
Verified Email
Yes
Verified Flair
No
Total Karma
432,840
Link Karma
34,836
Comment Karma
395,700
Profile updated: 1 day ago
Subreddit
Post Details
We try to extract some basic information from the post title. This is not
always successful or accurate, please use your best judgement and compare
these values to the post title and body for confirmation.
- Posted
- 8 months ago
- Reddit URL
- View post on reddit.com
- External URL
- reddit.com/r/OpenAI/comm...