Check out OPT-IML, but a lot of people said they didn't know:
– Flan-T5 checkpoints are publicly available (without requesting access)
– Flan-T5 11B outperforms OPT-IML on MMLU and Big-Bench Hard, despite being 10x more compute efficient Checkpoints: https://
huggingface.co/docs/transform
ers/model_doc/flan-t5
…
LLMS
-
Flan-T5 Outperforms OPT-IML with Greater Efficiency
By
–
-
Dot Product and Cosine Similarity in NLP Explained
By
–
In this post, @luis_likes_math defines two types of similarities for sentences: dot product similarity and cosine similarity. These similarities are very useful in determining if two sentences are similar or different. Learn more about similarity in NLP in this exciting post:
-
ChatGPT as synecdoche for all GPT-3 models bothers user
By
–
This bothers me too but I've given up on it. People outside LLMs use "ChatGPT" as a synecdoche for all GPT-3 models.
-
Microsoft Bing to incorporate OpenAI GPT-4, Semafor reports
By
–
Semafor reports Microsoft Bing will incorporate OpenAI's GPT-4: https://
semafor.com/article/02/01/
2023/chatgpt-is-about-to-get-even-better-and-microsofts-bing-could-win-big
… -
Constitutional AI: LLMs as Reward Model Annotators
By
–
Worth checking out constitutional AI from @AnthropicAI where LLMs can be the reward model annotators by prompting well. The choice then comes to paying for GPU or Phillipines.
-
NGram Example Selector for LangChain Prompt Engineering
By
–
ngram example selector Adding to the arsenal of existing example selectors, @seanspriggens added one based on ngram overlap. Select the closest examples by ngram overlap up to a threshold! Documentation: https://
langchain.readthedocs.io/en/latest/modu
les/prompts/examples/example_selectors.html#ngram-overlap-exampleselector
… -
LangChain from_string Method Simplifies LLMChain Construction
By
–
Usability is everything, and @brucehammer came up with a great idea for making it easier to construct LLMChains. A `from_string` method parses out the variables and auto-creates a prompt Documentation: https://
langchain.readthedocs.io/en/latest/modu
les/chains/generic/llm_chain.html#from-string
… -
Instruct Embeddings: Task-Tailored Text Embeddings from HuggingFace
By
–
@huggingface Instruct embeddings Instruct embeddings are from an instruction-finetuned embedding model that can generate text embeddings tailored to any task, *just by providing the task description* A joint effort by seanaedmiston and @EnoReyes https://
langchain.readthedocs.io/en/latest/modu
les/utils/combine_docs_examples/embeddings.html#instructembeddings
… -
LangChain 0.0.76 Release: Instruct Embeddings and New Features
By
–
Another big release! 0.0.76 main features: @huggingface Instruct embeddings (seanaedmiston, @EnoReyes
) ngram example selector (
@seanspriggens
) Other features include a new deployment template, easier way to construct LLMChain, and updates to PALChain Lets dive in -
MultiRay: Meta’s Efficient Platform for Large-Scale AI Models
By
–
MultiRay is Meta’s platform for efficiently running large-scale, state-of-the-art AI models. By converting input to an embedding that’s cached & shared between users, MultiRay greatly reduces complexity & energy use.