[adamwintle]

For most conversations I get: "I'm just a language model, so I can't help you with that.", "As a language model, I'm not able to assist you with that.", " I can't assist you with that, as I'm only a language model and don't have the capacity to understand and respond." — whereas ChatGPT gives very helpful replies...

[CrypticShift]

Someone should make a censorship/alignment (whatever you want to call it) benchmark for LLMs.

[thierrydamiba]

https://tatsu-lab.github.io/alpaca_eval/

Such a leaderboard exists, AlpacaEval Leaderboard ranks LLMs on the ability to follow user instructions.

[seydor]

i ve always had the opposite experience. Bard has not denied to help me writing a patent or writing a paper. ChatGPT denied both

[pram]

FWIW with the Assistants API you can instruct GPT to do anything you want and it won't have any "safety" denial messages.

[ikari_pl]

wow, the Google Assistant spirit continues!

[magicalhippo]

So Google's version of Clippy?