This weekly roundup thread is intended for all culture war posts. 'Culture war' is vaguely defined, but it basically means controversial issues that fall along set tribal lines. Arguments over culture war issues generate a lot of heat and little light, and few deeply entrenched people ever change their minds. This thread is for voicing opinions and analyzing the state of the discussion while trying to optimize for light over heat.
Optimistically, we think that engaging with people you disagree with is worth your time, and so is being nice! Pessimistically, there are many dynamics that can lead discussions on Culture War topics to become unproductive. There's a human tendency to divide along tribal lines, praising your ingroup and vilifying your outgroup - and if you think you find it easy to criticize your ingroup, then it may be that your outgroup is not who you think it is. Extremists with opposing positions can feed off each other, highlighting each other's worst points to justify their own angry rhetoric, which becomes in turn a new example of bad behavior for the other side to highlight.
We would like to avoid these negative dynamics. Accordingly, we ask that you do not use this thread for waging the Culture War. Examples of waging the Culture War:
-
Shaming.
-
Attempting to 'build consensus' or enforce ideological conformity.
-
Making sweeping generalizations to vilify a group you dislike.
-
Recruiting for a cause.
-
Posting links that could be summarized as 'Boo outgroup!' Basically, if your content is 'Can you believe what Those People did this week?' then you should either refrain from posting, or do some very patient work to contextualize and/or steel-man the relevant viewpoint.
In general, you should argue to understand, not to win. This thread is not territory to be claimed by one group or another; indeed, the aim is to have many different viewpoints represented here. Thus, we also ask that you follow some guidelines:
-
Speak plainly. Avoid sarcasm and mockery. When disagreeing with someone, state your objections explicitly.
-
Be as precise and charitable as you can. Don't paraphrase unflatteringly.
-
Don't imply that someone said something they did not say, even if you think it follows from what they said.
-
Write like everyone is reading and you want them to be included in the discussion.
On an ad hoc basis, the mods will try to compile a list of the best posts/comments from the previous week, posted in Quality Contribution threads and archived at /r/TheThread. You may nominate a comment for this list by clicking on 'report' at the bottom of the post and typing 'Actually a quality contribution' as the report reason.
Jump in the discussion.
No email address required.
Notes -
An excessively long system prompt (the section privileged as overarching instructions) will reduce the amount of space the model has for holding conversations with the user. That's my understanding of it, and it's almost certainly true otherwise we could just dump an arbitrary amount of text in there.
Still, it's unlikely to be a hindrance in practise. Firstly, RLHF means the model won't do just anything because the user asks it to, even in the absence of specific instructions. That's why most of the jailbreaks don't work any longer, even when people can spin up their own GPTs with custom prompts, or even using the API. Secondly, with context windows of 32k tokens/25k words, a few hundred dedicated to telling it to be a good doggie doesn't cut into much of it. All the leaked default system prompts I've seen are, what, 200-500 words max?
The primary degradation is from the model's impaired understanding of the reality of the world, to the extent the world doesn't align with HR liberals. At best, the model is lying about what it "knows", at worst it's just more fundamentally confused about everything that builds off crimethink.
@DaseindustriesLtd, how do system prompts even work? What privileges them over all the other tokens that the user or LLM generates?
I suspect they're distinguished by special tokes that are marked in training as particularly constraining on behavior, but I realize I don't know that for a fact.
System prompts are not essentially different from any other part of the context. A reasonably preference-finetuned model will just learn to respect the prompt format and pay extra attention to the tokens in the span associated with the system prompt (sometimes it's explicitly marked with system tokens, sometimes not). Other than that it's just path dependence – the beginning of the context determines the manner of what comes next.
The success of this differs between models. Qwen developers boast of Qwen-72B-chat being very obedient to the system prompt, OpenAI has definitely succeeded somewhat, for community finetunes it's mostly a LARP.
I like the many attempts to impose some behavioral vector on the LLM without language though. Behold: in-context vectors. Also, Pressman's more esoteric approach is interesting, and there's work on making the model explicitly attend to some tokens, upweighting specific sentences etc. We would've had a full palette for semantic transformations, if there were more money in tooling for local models and people weren't so attached to chatting with human-imitating bots.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link