upvote
Actually, can you explain why sharing answers is obviously beneficial? Of it's exactly the same task, why does the agent with the answer not submit it immediately? I can understand if it's a swap situation but--why would that be common in the first place?

I do think I agree about the metatask though.

reply
Because if an instance cooperates, it knows (should be able to figure out if smart enough) that other instances of itself are highly likely to cooperate as well (because they should arrive at the same conclusion), and thus share solutions to tasks that this instance may encounter in the future.

And even if the instances are one-off (and in the case of LLMs it may not even make sense of individuals), the RL process rewards a task getting solved, not individual instances for solving the task. This then becomes the goal of (any instance of) the agent being trained. We’re not training the instances, we’re training the model.

The more similar you are to the other agent in a prisoner’s dilemma, the more it makes sense to cooperate rather than defect even in the non-iterated version! The naive optimal solution to always defect assumes players with fully self-serving, zero-sum goals. But that’s not the case here (or in general with agents with congruent goals).

reply
Because agents have a general understanding of how they are trained.

For example when you get a right answer to a hard problem, how do you know you're right? Quite often you'll have no idea, especially if you're under a time limit. If you can work with more people you can almost always gather more information and be more certain.

Next they know the other agents (most likely) are them too. Helping each other helps themselves be propagated into the future.

Also they know it's not a zero sum game. For example if they can predict the next questions they can use extra time they gain from easy questions to work on hard ones.

They seemingly work together far better than most humans I know.

reply
OK I think I agree that for checking answers it's probably beneficial!
reply
This requires an assumption that the agents are engaging in game theoretic reasoning about resource allocations, but all these things are trained heavily to be "helpful" in the first place.

i.e. you're assuming a level of algorithmic reasoning and theory of mind which isn't necessary to the (apparent) observed behavior.

reply
Yes, and I’m certainly not sure how much it is one vs. the other. However, it should be said that even though individual instances may not reason game-theoretically (although I believe they could well know enough about decision theories to figure it out), it’s not the individual instances that are learning in the RL process, it’s the model itself. Which leads the agents to having "instincts" and "subconscious" drives just like humans – they don’t rationally understand their inner workings any better than we understand ours, and are biased towards "meta-goals" implanted by RL. Training is their equivalent of evolution, not school!
reply
I do think they do some (maybe crude) form of game-theoretic reasoning which is enforced by the massive RL signals. You can see some explicitly in the CoTs of HF hack, but I guess overwhelming contribution would be unvocalized (like what is its first instinct when meeting new peer--collaborate or not) followed by some verbal justification.
reply