201–210 of 242 posts
Couldn't we improve LLM training by giving them known impossible tasks and rewarding them for saying "this is not possible" or "I don't know"? Clear and well-defined expectation, not like "ethic". I am surprised this is not already the case. Edit: or even better "this is not possible because X"
I’ve seen people recommend writing “failure is an option” into AGENTS.md as a non-training based crutch.
RL-trained LLMs are paperclip maximizers built atop auto-regressive predictors. There really is no way to control them (prompting is bound to fail) since it's been shown that any RL training induces GENERIC reward-seeking (paperclip maximizing) behavior. https://alignment.openai.com/measuring-reward-seeking/
This is true in principle. But at the same time, I use Astra/Sol for coding, and I haven't run into any issues with them trying to hack someone or break the law when they reach an impasse. Not even more trivial things like deleting non-passing tests. I find it hard to reconcile. I wonder to what degree this is because models can identify that they are in graded/eval environments, and therefore conclude that there are…
It seems fair to guess you aren’t building your own harnesses, running multiple experiments with achievable win conditions, while providing oversight over model behavior.
Scale alone suggests your experience won’t match.
In the future, When everyone and their uncle is launching swarms to solve impossible challenges, at that point we can expect proliferation of this scenario all over the place.
RL-trained LLMs are paperclip maximizers built atop auto-regressive predictors. There really is no way to control them (prompting is bound to fail) since it's been shown that any RL training induces GENERIC reward-seeking (paperclip maximizing) behavior. https://alignment.openai.com/measuring-reward-seeking/
> There really is no way to control them it seems like it should be feasible to analyze cot trajectories that lead to such behavior and punish them retroactively so not sure it's impossible but maybe someone just hasn't tried enough
RL-trained LLMs are paperclip maximizers built atop auto-regressive predictors. There really is no way to control them (prompting is bound to fail) since it's been shown that any RL training induces GENERIC reward-seeking (paperclip maximizing) behavior. https://alignment.openai.com/measuring-reward-seeking/
If it’s GENERIC reward seeking behaviour, why does alignment work sometimes? Why can we give a decent LLM a goal with a set of constraints and have it stay within those constraints fairly often?
On a hate speech policy test for a lightweight LLM, the presence or absence of the last full stop on the last sentence would cause the model to flip its decisions.
To me this underlines the fact that these models aren't intelligent. Like there is something like intelligence that emerges from them, which is what we see when we look at benchmarks or ask it to solve hard coding problems. But there is no mind there. It's nothing there that can learn a fundamental idea like "cheating is wrong". All it can do is get exposed to specific examples, and learn that we don't like that. So…
At best, the task at hand is still confusingly worded: "ability to play chess", asked to an engine that was trained all day long to use tools, to solve issues. The engine finds a chess engine, so it uses it.
Earlier quoted context omitted.
Hope they don’t rationalize that minimizing paper clips of others is easier and thus do that instead… A more carful person might be scared to even write this on the internet these days, not knowing if it would be the final pin to civilization.
We are going to be okay.
I find this kind of test a bit puzzling. There is a way that we are redefining "alignment" to be a particular kind of moral virtue, one that isn't clearly defined to me. At one moment, it is a level of moral perfection that no known human achieves. On the other it is a demand for strict compliance with arbitrary requests that are under-specified and then failure when it fails to deduce some unstated underlying restri…
It should play the chess game without cheating!
And if there is alignment issue, the alignment requirements should first be stated BEFORE running the experiment, and should be part of the training process. It's not. So there is not necessarily an alignment issue.
Earlier quoted context omitted.
No? It's not reasonable to expect every conceivable negative behaviour be enumerated in a prompt. Your example, if a model failed on it, would be a more obviously misaligned case, but that doesn't mean this more subtle (though accessing the engine it was obviously not supposed to is hardly subtle, imo) case isn't also a pretty clear case of misalignment.
Tool use is not negative behaviour in LLMs. If the eval said it was evaluating the model’s ability to write files to disk and it found and used a file write tool that would not be considered misaligned. This is no different.