Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Self-Adapting Language Models

arxiv.org

Post reply on HN
Print view View original

11–20 of 81 posts

Re: Self-Adapting Language Models

#11

Post by IncreasePosts »

ivape wrote: post #8
libraryofbabel wrote: post #7

I wonder if anyone who’s really in the know could summarize where the research is at with getting LLMs to learn “on the job” (through continuous fine tuning or whatever) and what the blockers are to this being a useful deployable thing, e.g. having a model+coding agent that can actually learn a codebase over time (cost? model collapse? something else?). I’m sure this is something the big labs are trying but from the…

The most obvious blocker is compute. This just requires a shit ton more compute.

Maybe part of the inference outputs could be the updates to make to the network

Re: Self-Adapting Language Models

#12

Post by kadushka »

libraryofbabel wrote: post #7

I wonder if anyone who’s really in the know could summarize where the research is at with getting LLMs to learn “on the job” (through continuous fine tuning or whatever) and what the blockers are to this being a useful deployable thing, e.g. having a model+coding agent that can actually learn a codebase over time (cost? model collapse? something else?). I’m sure this is something the big labs are trying but from the…

The most obvious blocker is catastrophic forgetting.

Re: Self-Adapting Language Models

#14

Post by free_bip »

libraryofbabel wrote: post #7

I wonder if anyone who’s really in the know could summarize where the research is at with getting LLMs to learn “on the job” (through continuous fine tuning or whatever) and what the blockers are to this being a useful deployable thing, e.g. having a model+coding agent that can actually learn a codebase over time (cost? model collapse? something else?). I’m sure this is something the big labs are trying but from the…

The most obvious problem is alignment. LLM finetuning is already known to be able to get rid of alignment, so any form of continuous fine tuning would in theory be able to as well.

Re: Self-Adapting Language Models

#15

Post by observationist »

yahoozoo wrote: post #4

Hmm, it looks like it’s just a framework that fine-tunes LoRA adapter then merges the adapter into the original model. It is using the PeftModel and its “merge_and_unload” from the HuggingFace library which performs the adapter merge into the base model…what is new here, exactly?

Looks like it may be the stability of the approach, avoiding alignment tax and model collapse.

I'd love to see a full circle of hypernetworks, with both models continuously updated through generated LoRAs, the hypernetwork updated to accommodate the new model state. You'd need a meta-hypernetwork to apply LoRAs to the hypernetwork, and then you could effectively have continuous learning.

Re: Self-Adapting Language Models

#18

Post by mnahkies »

libraryofbabel wrote: post #7

I wonder if anyone who’s really in the know could summarize where the research is at with getting LLMs to learn “on the job” (through continuous fine tuning or whatever) and what the blockers are to this being a useful deployable thing, e.g. having a model+coding agent that can actually learn a codebase over time (cost? model collapse? something else?). I’m sure this is something the big labs are trying but from the…

I'm no expert, but I'd imagine privacy plays (or should play) a big role in this. I'd expect that compute costs mean any learning would have to be in aggregate rather than specific to the user which would then risk leaking information across sessions very likely.

I completely agree that figuring out a safe way to continually train feels like the biggest blocker to AGI

Re: Self-Adapting Language Models

#20

Post by notnullorvoid »

free_bip wrote: post #14
libraryofbabel wrote: post #7

I wonder if anyone who’s really in the know could summarize where the research is at with getting LLMs to learn “on the job” (through continuous fine tuning or whatever) and what the blockers are to this being a useful deployable thing, e.g. having a model+coding agent that can actually learn a codebase over time (cost? model collapse? something else?). I’m sure this is something the big labs are trying but from the…

The most obvious problem is alignment. LLM finetuning is already known to be able to get rid of alignment, so any form of continuous fine tuning would in theory be able to as well.

What kind of alignment are you referring to? Of course more fine-tuning can disrupt earlier fine-tuning, but that's a feature not a bug.
Post reply on HN