3 ms·

With Qwen I went as stupid as I could: please provide the bounding box metadata for pytesseract for the above image. And it spat it out.

VladVladikoff 1y ago
It’s funny that many of us say please. I don’t think it impacts the output, but it also feels wrong without it sometimes.
- wongarsu 1y ago
  Depends on the model, but e.g. [1] found many models perform better if you are more polite. Though interestingly being rude can also sometimes improve performance at the cost of higher bias Intuitively it makes sense. The best sources tend to be either of moderately high politeness (professional language) or 4chan-like (rude, biased but honest) 1: https://arxiv.org/pdf/2402.14531 https://arxiv.org/pdf/2402.14531
  - arcanemachiner 1y ago
    When I want an LLM to be be brief, I will say things like "be brief", "don't ramble", etc. When that fails, "shut the fuck up" always seems to do the trick.
    - richardlblair 1y ago
      I ripped into cursor today. It didn't change anything but I felt better lmao
  - entropie 1y ago
    Bevore GPT5 was released I already had the feeling like the webui response was declining and I started to try to get more out of the responses and dissing it and saying how useless their response was did actually improve the output (I think).
- indigoabstract 1y ago
  The way I think of it, talking to an LLM is a bit like talking to myself or listening to an echo, since what I get back depends only on what I put in. If it senses that I'm frustrated, it will be inclined to make even more stuff up in an attempt to appease me, so that gets me nowhere. I've found it more useful to keep it polite and "professional" and restart the conversation if we've begun going around in circles. And besides, if I make a habit of behaving badly with LLMs, there's a good chance that I'll do it without thinking at some point and get in trouble.
- dabockster 1y ago
  It's a good habit to build now in case AGI actually happens out of the blue.

VladVladikoff 1y ago

It’s funny that many of us say please. I don’t think it impacts the output, but it also feels wrong without it sometimes.

wongarsu 1y ago
Depends on the model, but e.g. [1] found many models perform better if you are more polite. Though interestingly being rude can also sometimes improve performance at the cost of higher bias Intuitively it makes sense. The best sources tend to be either of moderately high politeness (professional language) or 4chan-like (rude, biased but honest) 1: https://arxiv.org/pdf/2402.14531 https://arxiv.org/pdf/2402.14531
- arcanemachiner 1y ago
  When I want an LLM to be be brief, I will say things like "be brief", "don't ramble", etc. When that fails, "shut the fuck up" always seems to do the trick.
  - richardlblair 1y ago
    I ripped into cursor today. It didn't change anything but I felt better lmao
- entropie 1y ago
  Bevore GPT5 was released I already had the feeling like the webui response was declining and I started to try to get more out of the responses and dissing it and saying how useless their response was did actually improve the output (I think).
indigoabstract 1y ago
The way I think of it, talking to an LLM is a bit like talking to myself or listening to an echo, since what I get back depends only on what I put in. If it senses that I'm frustrated, it will be inclined to make even more stuff up in an attempt to appease me, so that gets me nowhere. I've found it more useful to keep it polite and "professional" and restart the conversation if we've begun going around in circles. And besides, if I make a habit of behaving badly with LLMs, there's a good chance that I'll do it without thinking at some point and get in trouble.
dabockster 1y ago
It's a good habit to build now in case AGI actually happens out of the blue.