HN Explorer

Ask HN: How do you make LLM generated text believable?

3 points by StahlGuo ↗ HN
As a graudate student who just working as management & IR, I use LLM to do daily jobs , including weekly briefing and But AI generated report looks too good to be checked, and hallucination can't be terminated. But Boss and SEC can not tolerate any number mistakes. So, can you tell me, how do you solve this problem?

8 comments

[ 2.3 ms ] story [ 23.2 ms ] thread
bigyabai
...do you actually send the SEC LLM-generated documents?

Do you understand the liability you're assuming, doing that?

StahlGuo
Oh no I am not a attorney, but I need to analyze the market, 20-F and 10-K recurringly and make strategy suggestions, and so yes I use LLM to do business brief.
0x1d7
> So, can you tell me, how do you solve this problem?

Write it yourself.

dabinat
If you were writing it yourself you would proofread and double-check everything. Why do you think that’s no longer required?
evil-olive
> But AI generated report looks too good to be checked

have you considered...checking it anyway, no matter how good it looks?

StahlGuo
I don't mean never check it by myself, but I want to discuss about a methodology to ensure AI generated text has the same auditability and tracebility like human wrote text.
StahlGuo
I am sorry that I did not explain my question explicitly. I write recurring weekly briefings for internal use, such as market insights and industry news. I’m not trying to make AI-generated text “believable”. I’m asking almost the opposite question: when an AI generated text is fluent enough to hide mistakes, how do human check how to systematically check numbers, dates, cites and judgements?