让ChatGPT更加诚实

Just a heads up, if you want honest AI stuff, the best approach is to combine multiple tools and compare answers. Don’t rely on just one output.

Honestly, sometimes I just chat with it casually first to warm it up, then get serious after. Weird but it seems to respond better that way.

I guess it also depends on your expectations. Sometimes you just gotta remember it’s a language model, not a human with feelings or biases.

Sometimes I try to remind it that I’m just looking for facts, not trying to argue or anything. That helps it focus on honesty over fluff answers.

Maybe if they add a feature where it can say 'I'm not sure' more often, it’d feel more honest.

You can also check ai-u.com for new or trending tools that might have better honesty features or more transparent AI models.

One thing I found is that ChatGPT sometimes tries not to offend or upset, so it avoids blunt honesty. So if you want more straightforward answers, you might have to explicitly say it's okay to be frank.

Honestly, sometimes it just makes stuff up when it doesn't know the answer, which is frustrating. I wish there was a better way to get it to admit "I don't know" instead of guessing.

I sometimes phrase stuff like ‘be brutally honest’ and it seems to loosen up the responses a bit. Works better than just ‘be honest’ imo.

I tried asking it to list pros and cons on stuff, and that usually brings out more balanced and honest info.

I think the way ChatGPT is trained, it just wants to be helpful and polite, so it avoids anything that could come off as rude or too honest in a harsh way.

For me, honesty also means it doesn’t make up crazy facts. So if you catch it doing that, just ask it to verify or cite sources.

I've noticed that sometimes phrasing your questions more specifically helps. If you ask vague stuff, it kinda hedges. But if you get real clear, it tends to be more direct.

I’ve noticed that if you ask the same question a few times, you get slightly different answers. Maybe that randomness affects honesty too?

It's weird how sometimes it gives a perfect honest answer, and then other times it kinda dances around the topic. No idea why the inconsistency.

I’m curious if future models will just come with an honesty mode or a truth rating on answers, that’d be sweet.