zlacker

[parent] [thread] 3 comments
1. xiansh+(OP)[view] [source] 2025-05-14 20:36:58
Amusingly, about 90% of my rat's-nest problems with Sonnet 3.7 are solved by simply appending a few words to the end of the prompt:

"write minimum code required"

It's not even that sensitive to the wording - "be terse" or "make minimal changes" amount to the same thing - but the resulting code will often be at least 50% shorter than the un-guided version.

replies(2): >>panstr+J2 >>theshr+Tn1
2. panstr+J2[view] [source] 2025-05-14 20:50:48
>>xiansh+(OP)
Well, the article mentions that this reduces accuracy. Do you hit that problem often then?
replies(1): >>SteveM+Jo
◧◩
3. SteveM+Jo[view] [source] [discussion] 2025-05-14 23:55:19
>>panstr+J2
The study the article cited is specifically about when asking the LLMs about misinformation. I think on coding tasks and such shorter answers are usually more accurate.
4. theshr+Tn1[view] [source] 2025-05-15 12:17:49
>>xiansh+(OP)
Gemini on the other hand has a tendency for super-defensive coding.

It'll check _EVERY_ edge case separately, even in situations where it will never ever happen and if it does, it's a NOP anyway.

[go to top]