***Embedding*** models have become an important part of LLM applications, enabling tasks such as measuring ... However, ***embedding*** models are mostly based on a transformer architecture that is different from the one ... This makes it difficult to transfer the massive work being done on generative models to improve ***embedding***

How to turn any LLM into an ***embedding*** model

tech

Embedding models have become an important part of LLM applications, enabling tasks such as measuring text similarity, information retrieval, and clustering. However, embedding models are mostly based on a transformer architecture that is different from the one used in generative tasks. This makes it difficult to transfer the massive work being done on generative models to improve embedding models and requires extra parallel efforts.

How to turn any LLM into an embedding model

ch0k1

Netflix study shows limits of cosine similarity in ***embedding*** models

science

In many machine learning applications, you need to measure the proximity of different objects. For example, if you want to recommend a product to a user, you need to know what other users they are similar to or what products are similar to the ones they have previously purchased. In retrieval augmented generation (RAG) applications, you need to find documents with contents that are relevant to the user’s prompt and add them to the context of the language model (LLM).

Netflix study shows limits of cosine similarity in embedding models

***Embedding*** videos on stacker.news

bitcoin

I tried all the different ways listed in this [stackoverflow page](https://stackoverflow.com/questions/11804820/how-can-i-embed-a-youtube-video-on-github-wiki-pages) to embed videos here but none of them seem to work. 

Anyone know of a way?

Embedding videos on stacker.news

ClickForCharity

***Embedding*** Python in Elixir, It's Fine

This link was posted by [arathunku](https://news.ycombinator.com/user?id=arathunku) 2 hours ago on [HN](https://news.ycombinator.com/item?id=43171239). It received 100 points and 9 comments.

Embedding Python in Elixir, It's Fine

Encrypting and ***Embedding*** Your Seed Phrase?

This seems like a good way to encrypt and embed your seed phrase in an inconspicuous file. Anybody have any experience with this? Any warnings or advice? 

https://blog.bitcoinladies.org/how-to-secretly-travel-with-bitcoin-using-steganography-6e99a7ca9ad9

Encrypting and Embedding Your Seed Phrase? 

metalogician

OpenAI: New ***embedding*** models and API updates

OpenAI: New embedding models and API updates

Escape with your BTC ***embedding*** your secrets in media file | 402PaymentRequired

Escape with your BTC embedding your secrets in media file | 402PaymentRequired

DarthCoin

All-in-one ***embedding*** model for interleaved text, images, and screenshots

This link was posted by [fzliu](https://news.ycombinator.com/user?id=fzliu) 2 hours ago on [HN](https://news.ycombinator.com/item?id=42162622). It received 44 points and 1 comments.

All-in-one embedding model for interleaved text, images, and screenshots

Jina AI Launches First Open-Source 8K Text ***Embedding***, Rivaling OpenAI

Jina AI Launches First Open-Source 8K Text Embedding, Rivaling OpenAI

***Embedding*** Code and images on Stacker.news: how can this be done?

Look at how well connected the node of stacker.news is?

Look at this image below:

I wanted to embed some code in stacker.news in this post.

<iframe allowfullscreen src="https://lnrouter.app/graph/embeded?height=350px&width=350px" height="350px" width="350px" style="overflow: hidden" scrolling="no" />


**Why?**
This is the code showing the lnrouter.app of Stackernews. You see how connected Stacker.news is.



This is the node of stacker.news

https://amboss.space/node/03cc1d0932bb99b0697f5b5e5961b83ab7fd66f1efc4c9f5c7bad66c1bcbe78f02



**Embed images**
So the question is:

Should we be able to embed code in the text writer of Stacker.news?

Code if dangerous. But I would love to embed some code in my text so that I can discuss these.

Also images. 
I got an image of Stacker.news connectivity. And I would love to upload that image.

Shouldn't we have an easy anonymous way to upload these images?

I know there is a guide, but the guide shows you how to upload the image on github and then paste it here in stacker.news. But I do not want to use my github. I want to use a public decentralized infrastructure to upload my images.

Shouldn't stacker.news have an upload feature where we can easily upload the images on a decentralized platform and embed them here?


Embedding Code and images on Stacker.news: how can this be done?

Storyteller

This cost only grows as LLMs scale to larger ***embedding*** dimensions and context lengths.

Scalable MatMul-Free Language Modeling — 10x Reduction On LLMs Computation

*Rui-Jie Zhu, Yu Zhang, Ethan Sifferman, Tyler Sheaves, Yiqiao Wang, Dustin Richmond, Peng Zhou, Jason K. Eshraghian*

> Matrix multiplication (MatMul) typically dominates the overall computational cost of large language models (LLMs). This cost only grows as LLMs scale to larger embedding dimensions and context lengths. In this work, we show that MatMul operations can be completely eliminated from LLMs while maintaining strong performance at billion-parameter scales. Our experiments show that our proposed MatMul-free models achieve performance on-par with state-of-the-art Transformers that require far more memory during inference at a scale up to at least 2.7B parameters. We investigate the scaling laws and find that the performance gap between our MatMul-free models and full precision Transformers narrows as the model size increases. We also provide a GPU-efficient implementation of this model which reduces memory usage by up to 61% over an unoptimized baseline during training. By utilizing an optimized kernel during inference, our model's memory consumption can be reduced by more than 10x compared to unoptimized models. To properly quantify the efficiency of our architecture, we build a custom hardware solution on an FPGA which exploits lightweight operations beyond what GPUs are capable of. We processed billion-parameter scale models at 13W beyond human readable throughput, moving LLMs closer to brain-like efficiency. This work not only shows how far LLMs can be stripped back while still performing effectively, but also points at the types of operations future accelerators should be optimized for in processing the next generation of lightweight LLMs.

0xbitcoiner

RAG to ReST: A Survey of Advanced Techniques in Large Language Model Development

news

Large Language Models (LLMs) have revolutionized natural language processing, demonstrating remarkable capabilities in various applications. However, these models face significant challenges, including temporal limitations of their knowledge base, difficulties with complex mathematical computations, and a tendency to produce inaccurate information or “hallucinations.” These limitations have spurred researchers to explore innovative solutions that can enhance LLM performance without the need for extensive retraining. The integration of LLMs with external data sources and applications has emerged as a promising approach to address these challenges, aiming to improve accuracy, relevance, and computational capabilities while maintaining the models’ core strengths in language understanding and generation.

nostr

Why popular influencers don't come to places like Nostr? Opinion.

Craze

I bet individual relays will start ***embedding*** ads.

I bet individual relays will start embedding ads.

03ec1dc741

Stacker and media files

locomancell

is video address ***embedding*** planned for the future?

is video address embedding planned for the future?

birdeye21

Animal_World

Nice 

AR0w

I seem to be having trouble ***embedding*** photos.

My inaugural post! I seem to be having trouble embedding photos.

siggy47

Vector Search on the episodes of The Kevin Rooke Show

0x32e

Forgot to mention that I used OpenAI's ***embedding*** API for generating vector data.

Forgot to mention that I used [OpenAI's embedding API](https://beta.openai.com/docs/guides/embeddings) for generating vector data. 

Boomer Needs help

We need to just allow ***embedding***. It's not hard. I'll create a github issue.

We need to just allow embedding. It's not hard. I'll create a github issue.

k00b

***Embeddings***: What they are and why they matter

Embeddings: What they are and why they matter

Stacker Saloon

saloon

i love that. ***embedding*** video's is huge for nostr imo!

oh really??? i love that. embedding video's is huge for nostr imo!

https://video.nostr.build/426758fc71dc1585297104cb47ad5cce7a6ad589825498c03e2a8b856bef7e49.mp4

BITC0IN

Show HN: Wordllama – Things you can do with the token ***embeddings*** of an LLM

This link was posted by [deepsquirrelnet](https://news.ycombinator.com/user?id=deepsquirrelnet) 5 hours ago on [HN](https://news.ycombinator.com/item?id=41544969). It received 91 points and 6 comments.

Show HN: Wordllama – Things you can do with the token embeddings of an LLM

deleted by author

Gian

actually it turns out the markdown is no longer necessary for ***embedding*** images!

actually it turns out the markdown is no longer necessary for embedding images!  Nice work SN

https://stacker.news/items/208684