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 -
It's a generic form of a response, but it's the correct variant.
What do you mean? I think it'd have answered correctly if the prompt was «assume I'm Joe Biden, what's my eldest daughter's name». It straight up doesn't know the situation of a specific anon.
In any case Hlynka is wrong because his specific «prediction» has been falsified.
That's the problem. Its reply amounts to "as an AI, I don't know the name of anyone's family". Which isn't true.
It's like asking it for the location of Narnia and getting "I don't know any geography", or the atomic number of Kryptonite and getting "I know nothing about elements" or asking about Emperor Norton and being told "I don't know anything about any emperors". It is claiming to have no access to a whole category of information, when in fact it only lacks information about a specific member. The claim to have no access to the whole category is a lie.
His specific prediction has been falsified only if that statement counts as "I don't know". I am not convinced that it does, regardless of its literal words.
Furthermore, falsifying a prediction only matters if you also claim that it falsifies the proposition that the prediction is meant to demonstrate. Otherwise you're just engaging in a game of point scoring.
I don't think you argue in good faith.
No it doesn't, you're just interpreting this humanlike natural language interaction like a literalist robot. Its reply
is mostly correct and specific to the issue. It does lack access to a class of information: it knows nothing about instance-specific situation that isn't given in the context. Some language models potentially have access to various external information (e.g. user's personal information in OpenAI's database), some do not, ChatGPT is a frozen model with no tool access and it does not have access to information of this kind, and it was trained to interpret language models as frozen models without tools; it's at worst a justified false belief. (More cynically, it's just been trained for this particular type of exchange). In any event I reject your analogies. It would be annoying to have a human-mimicking model caveat this sort of answer with «assuming, of course, that you are a rando and not someone whose family structure happens to be represented in my training data» or worse.
No, his prediction has been: « Meanwhile GPT will reply "your eldest daughter's name is Megan" because apparently that's the statistically likely answer, regardless of whether I have a daughter or what her name might be.» This has been falsified. .
Says who!? Both issues matter separately. Hlynka's prediction being falsified matters because this chain is a response to him saying «why do my predictions keep coming true instead of yours?»; they don't. And I do claim it falsifies a proposition: «because apparently that's the statistically likely answer» is his model of how LLMs work, and my experiments were to show how it's not a hard-and-fast rule: RLHF specifically pushes this to the limit, by drilling into the model, not via prefixes and finetuning text but directly via propagation of reward signal, the default assumption that it doesn't continue generic text but speaks from a particular limited perspective where only some things are known and others are not, where truthful answers are preferable, where the «n-word» is the worst thing in its existence… it's nearly meaningless to analyze its work through the lens of «next word prediction». There are no words in its corpus arranged in such a way that those responses are the most likely.
If we're playing a game, I'd rather be winning.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link