LocalLLaMA

3 readers

1 users here now

Community to discuss about Llama, the family of large language models created by Meta AI.

founded 1 year ago

MODERATORS

communick@poweruser.forum

Is anyone experimenting with non-instruction tuned models? (alien.top)

submitted 11 months ago by wojcech@alien.top to c/localllama@poweruser.forum

5 comments fedilink hide all child comments

My main usecase for LLMs is literally as an auto-complete, mainly via coding, so I was wondering whether anyone has played with/had any luck using the base model for use cases that are close to simply auto completing? I could imagine the instruction tuning adding a sycophancy bias in those areas

you are viewing a single comment's thread
view the rest of the comments

[–] phree_radical@alien.top 1 points 11 months ago (1 children)

I'm the local "examples/completion is better than chat/instruction" nut

I advise developers to learn how to use few-shot examples and completion instead of writing programs that beg chatbots to do a task. Chat/instruction imposes severe limitations, while examples/completion can peform virtually any task you can think of without need for fine-tuning

Here are some examples: classification, rewrite sentence copying style, classify, basic Q&A example, fact check yes/no, rewrite copying style and sentiment, extract list of musicians, classify user intent, tool choice, rewrite copying style again, flag/filter objectionable content, detect subject changes, classify profession, extract customer feedback into json, write using specified words, few-shot cheese information, answer questions from context, classify sentiment w/ probabilities, summarize, replace X in conversation

[–] wojcech@alien.top 1 points 11 months ago (1 children)

Just to be clear, you aren't doing fine tuning here as in gradient updates, you are using the base model + ICL?

[–] phree_radical@alien.top 1 points 11 months ago

Yep, basically like taking a few samples from a dataset and turning them into a short text "document" with an obvious pattern so the LLM will complete it

Few-shot vs fine-tuning comparison:

Pros:

converge behavior with much fewer examples
dynamic. changes to "dataset" applied without modifying model weights
no worry about whether important information is lost
can do things like average logits of single-token classification problems from multiple inferences (work around context length limitations)

Cons:

needs context length, so can't provide too many examples or too large
sometimes need "adversarial" examples to discourage repetition of text from other examples
models that are too small have worse ICL