What he says is that they can solve blocks world problems if you take them by the hand and show them how to do it, for every single problem you wan them to solve. Which is not very useful at all. He's arguing that this is what CoT achieves: it takes the LLM by the hand and introduces domain knowledge that the user has in every step of the way, so that it's not the LLM that's solving anything but the person using the LLM with CoT.
Do read the linked paper, it goes over this in more detail. It's a bit unfortunate that Rao chooses to communicate through the very noisy medium of twitter, but that's the internet.
The important step is that they demonstrate that there are no architectural limits that keep the LLM from acting in these domains, "only" knowledge/planning ones. Once we have a big enough dataset of CoT prompts, the model "just" has to generalize CoT prompting, not CoT following. It decomposes the problem into two halves, and demonstrates that if one half (instruction generation) is provided, the other (instruction following) becomes very tractable.
As to blocks world planning in particular, LLMs already have plenty of examples of block stacking problems - those are the standard motivational experiment in planning papers, like solving mazes is for Reinforcement Learning. Google returns 10 pages of results for "blocks world planning". If LLMs were capable of generalising as well as you expect they will one day, they should already be capable of solving block stacking problems without CoT and with no guidance.
I think mostly, LLMs at the moment are incredibly uneven. LLM assistants can pull obscure knowledge out of nowhere one second and fail extremely basic reasoning the next. So just because some example happens a lot in the source material doesn't mean the LLM can learn it. IMO, that CoT works at all is more down to luck of the training set than any inherent capability of the LLM.
Probably he's better tuned in to the tone of the discourse than me but I feel like if you showed CoT helped with Blocksworld then that would be a huge and triumphant result, rather than something taken for granted that we must loudly disprove!
>> And yet the practitioners of CoT swear that any and every problem can be solved with LLM by giving it a bit of a CoT help.
For example, see this arxiv paper:
Generalized Planning in PDDL Domains with Pretrained Large Language Models
https://arxiv.org/abs/2305.11014
Where the authors conclude:
In this work, we showed that GPT-4 with CoT summarization and automated debugging is a surprisingly strong generalized planner in PDDL domains.
The author of the tweet is an expert on planning and he's responding to that kind of thing.
> Where the authors conclude:
> ...is a surprisingly strong generalized planner
Wtf is going on here?
> Ever since I came across the Chain of Though (CoT) for LLMs paper, I wondered how it can possibly make sense given that there is little reason to believe that LLMs can follow procedures and unroll them for the current problem at hand. (After all, if they can do that, they should be able to do verification and general reasoning too--and we know they suck at those).
I don't follow. Why would procedure following imply general reasoning?