Saahil Ognawala
@saahilognawala.bsky.social
Head of Product @jina-ai.bsky.social AI, software, security, product management
Embedding models become "blind" beyond 4K tokens in context length. Building on the NoLIMA paper, our experiments show that for needle-in-a-haystack tasks, performance of embedding models drops to near-random chance with long contexts—even with exact keyword matches 🤔 🧵
March 7, 2025 at 9:28 AM
Embedding models become "blind" beyond 4K tokens in context length. Building on the NoLIMA paper, our experiments show that for needle-in-a-haystack tasks, performance of embedding models drops to near-random chance with long contexts—even with exact keyword matches 🤔 🧵
Reposted by Saahil Ognawala
I applied LLMs for query expansion and we wrote this article:
It sees to work out-of-the-box and generally boost the performance of embedding models. However, it requires more latency. Would be interesting to see more about this.
📃: jina.ai/news/query-e...
🛠️: github.com/jina-ai/llm-...
It sees to work out-of-the-box and generally boost the performance of embedding models. However, it requires more latency. Would be interesting to see more about this.
📃: jina.ai/news/query-e...
🛠️: github.com/jina-ai/llm-...
Query Expansion with LLMs: Searching Better by Saying More
Search has changed a lot since embedding models were introduced. Is there still a role for lexical techniques like query expansion in AI? We think so.
jina.ai
February 18, 2025 at 8:29 AM
I applied LLMs for query expansion and we wrote this article:
It sees to work out-of-the-box and generally boost the performance of embedding models. However, it requires more latency. Would be interesting to see more about this.
📃: jina.ai/news/query-e...
🛠️: github.com/jina-ai/llm-...
It sees to work out-of-the-box and generally boost the performance of embedding models. However, it requires more latency. Would be interesting to see more about this.
📃: jina.ai/news/query-e...
🛠️: github.com/jina-ai/llm-...
It's time for enterprise SaaS companies to seriously rethink what it is that their customers were paying them for, and for what will they now? It's no longer enough to operate a walled-garden service (rule-based ER mgmt) when "comprehension" of multimodal data is being commoditized by large models.
February 16, 2025 at 8:27 AM
It's time for enterprise SaaS companies to seriously rethink what it is that their customers were paying them for, and for what will they now? It's no longer enough to operate a walled-garden service (rule-based ER mgmt) when "comprehension" of multimodal data is being commoditized by large models.
I don't get it - are influencers really writing obituaries for nvda because there's a really good new model?? Make it make sense. It's obvs great that r1 training recipe is kinda open (which is btw charitable, cuz it's not even that open, there's nothing in there about training data)
January 27, 2025 at 8:03 PM
I don't get it - are influencers really writing obituaries for nvda because there's a really good new model?? Make it make sense. It's obvs great that r1 training recipe is kinda open (which is btw charitable, cuz it's not even that open, there's nothing in there about training data)
Reposted by Saahil Ognawala
Our submission to ECIR 2025 on jina-embeddings-v3 has been accepted! 🎉
At the ECIR Industry Day my colleague @str-saba.bsky.social presents how we train the latest version of our text embedding model.
More details on ECIR: ecir2025.eu
More details about the model: arxiv.org/abs/2409.10173
At the ECIR Industry Day my colleague @str-saba.bsky.social presents how we train the latest version of our text embedding model.
More details on ECIR: ecir2025.eu
More details about the model: arxiv.org/abs/2409.10173
47th EUROPEAN CONFERENCE ON INFORMATION RETRIEVAL – 47th EUROPEAN CONFERENCE ON INFORMATION RETRIEVAL
ecir2025.eu
December 16, 2024 at 4:18 PM
Our submission to ECIR 2025 on jina-embeddings-v3 has been accepted! 🎉
At the ECIR Industry Day my colleague @str-saba.bsky.social presents how we train the latest version of our text embedding model.
More details on ECIR: ecir2025.eu
More details about the model: arxiv.org/abs/2409.10173
At the ECIR Industry Day my colleague @str-saba.bsky.social presents how we train the latest version of our text embedding model.
More details on ECIR: ecir2025.eu
More details about the model: arxiv.org/abs/2409.10173
After having set up Claude with Brave and knowledge graph MCPs, I've tbh not felt the need to open chatgpt or ppl in a while.
December 12, 2024 at 6:40 AM
After having set up Claude with Brave and knowledge graph MCPs, I've tbh not felt the need to open chatgpt or ppl in a while.
Reposted by Saahil Ognawala
I can now run a GPT-4 class model on my laptop
(The exact same laptop that could just about run a GPT-3 class model 20 months ago)
The new Llama 3.3 70B is a striking example of the huge efficiency gains we've seen in the last two years
simonwillison.net/2024/Dec/9/l...
(The exact same laptop that could just about run a GPT-3 class model 20 months ago)
The new Llama 3.3 70B is a striking example of the huge efficiency gains we've seen in the last two years
simonwillison.net/2024/Dec/9/l...
I can now run a GPT-4 class model on my laptop
Meta’s new Llama 3.3 70B is a genuinely GPT-4 class Large Language Model that runs on my laptop. Just 20 months ago I was amazed to see something that felt …
simonwillison.net
December 9, 2024 at 3:19 PM
I can now run a GPT-4 class model on my laptop
(The exact same laptop that could just about run a GPT-3 class model 20 months ago)
The new Llama 3.3 70B is a striking example of the huge efficiency gains we've seen in the last two years
simonwillison.net/2024/Dec/9/l...
(The exact same laptop that could just about run a GPT-3 class model 20 months ago)
The new Llama 3.3 70B is a striking example of the huge efficiency gains we've seen in the last two years
simonwillison.net/2024/Dec/9/l...
Reposted by Saahil Ognawala
One year ago, we released the first OS embedding model for 8192 tokens. Many suspected it to be not useful and chunking to be better than a single vector. I run many experiments to explain, when to use what and we summarized the findings in this article
t.co/BLC3WTU3LP
t.co/BLC3WTU3LP
https://jina.ai/news/still-need-chunking-when-long-context-models-can-do-it-all/
t.co
December 5, 2024 at 8:49 AM
One year ago, we released the first OS embedding model for 8192 tokens. Many suspected it to be not useful and chunking to be better than a single vector. I run many experiments to explain, when to use what and we summarized the findings in this article
t.co/BLC3WTU3LP
t.co/BLC3WTU3LP
Thanks for reminding me, XAI, because I had forgotten to tell everyone that Jina Embeddings API is also OpenAI compatible 🤭
December 4, 2024 at 2:00 PM
Thanks for reminding me, XAI, because I had forgotten to tell everyone that Jina Embeddings API is also OpenAI compatible 🤭
Got a bit into a rabbit hole of image modelling, after talking to a colleague last week. While Jina CLIP models employ EVA-02 as image tower, I wanted to know if there's an equivalent to masked "patch prediction" for images, like MLM in pre-trained text models arxiv.org/abs/2303.11331 (1/n)
EVA-02: A Visual Representation for Neon Genesis
We launch EVA-02, a next-generation Transformer-based visual representation pre-trained to reconstruct strong and robust language-aligned vision features via masked image modeling. With an updated pla...
arxiv.org
December 3, 2024 at 2:12 PM
Got a bit into a rabbit hole of image modelling, after talking to a colleague last week. While Jina CLIP models employ EVA-02 as image tower, I wanted to know if there's an equivalent to masked "patch prediction" for images, like MLM in pre-trained text models arxiv.org/abs/2303.11331 (1/n)
I've had a funny year really. For most of it, weeks flew past me so quickly because, well AI and it's 2024, but at least weekends were slow, inside or at a playground. But ever since I restarted seeing friends actively since the birth of my second child, the weekends seem to whizz right by too.
December 1, 2024 at 8:33 PM
I've had a funny year really. For most of it, weeks flew past me so quickly because, well AI and it's 2024, but at least weekends were slow, inside or at a playground. But ever since I restarted seeing friends actively since the birth of my second child, the weekends seem to whizz right by too.
Reposted by Saahil Ognawala
And the idea that "is it public" is all that matters and not what you DO with people's content, is absurd.
Like you cannot possibly suggest with a straight face that this example of using transgender YouTubers' videos to train facial recognition is 100% fine.
www.theverge.com/2017/8/22/16...
Like you cannot possibly suggest with a straight face that this example of using transgender YouTubers' videos to train facial recognition is 100% fine.
www.theverge.com/2017/8/22/16...
Transgender YouTubers had their videos grabbed to train facial recognition software
In the race to train AI, researchers are taking data first and asking questions later
www.theverge.com
November 27, 2024 at 3:56 PM
And the idea that "is it public" is all that matters and not what you DO with people's content, is absurd.
Like you cannot possibly suggest with a straight face that this example of using transgender YouTubers' videos to train facial recognition is 100% fine.
www.theverge.com/2017/8/22/16...
Like you cannot possibly suggest with a straight face that this example of using transgender YouTubers' videos to train facial recognition is 100% fine.
www.theverge.com/2017/8/22/16...
Had a lot of fun recording this one with @nicolay.fyi. All things considered, in the time between recording it and now I haven't seen many left-field developments that have changed my mind about evals in RAG.
open.spotify.com/episode/5bzb...
open.spotify.com/episode/5bzb...
RAG's Biggest Problems & How to Fix It (ft. Synthetic Data) | S2 E16
open.spotify.com
November 28, 2024 at 1:17 PM
Had a lot of fun recording this one with @nicolay.fyi. All things considered, in the time between recording it and now I haven't seen many left-field developments that have changed my mind about evals in RAG.
open.spotify.com/episode/5bzb...
open.spotify.com/episode/5bzb...
Reposted by Saahil Ognawala
AI is good at pricing, so when GPT-4 was asked to help merchants maximize profits - and it did exactly that by secretly coordinating with other AIs to keep prices high!
So... aligned for whom? Merchants? Consumers? Society? The results we get depend on how we define 'help' arxiv.org/abs/2404.00806
So... aligned for whom? Merchants? Consumers? Society? The results we get depend on how we define 'help' arxiv.org/abs/2404.00806
November 28, 2024 at 5:15 AM
AI is good at pricing, so when GPT-4 was asked to help merchants maximize profits - and it did exactly that by secretly coordinating with other AIs to keep prices high!
So... aligned for whom? Merchants? Consumers? Society? The results we get depend on how we define 'help' arxiv.org/abs/2404.00806
So... aligned for whom? Merchants? Consumers? Society? The results we get depend on how we define 'help' arxiv.org/abs/2404.00806
Reposted by Saahil Ognawala
The authors of ColPali trained a retrieval model based on SmolVLM 🤠 TLDR;
- ColSmolVLM performs better than ColPali and DSE-Qwen2 on all English tasks
- ColSmolVLM is more memory efficient than ColQwen2 💗
Find the model here huggingface.co/vidore/colsm...
- ColSmolVLM performs better than ColPali and DSE-Qwen2 on all English tasks
- ColSmolVLM is more memory efficient than ColQwen2 💗
Find the model here huggingface.co/vidore/colsm...
November 27, 2024 at 2:10 PM
The authors of ColPali trained a retrieval model based on SmolVLM 🤠 TLDR;
- ColSmolVLM performs better than ColPali and DSE-Qwen2 on all English tasks
- ColSmolVLM is more memory efficient than ColQwen2 💗
Find the model here huggingface.co/vidore/colsm...
- ColSmolVLM performs better than ColPali and DSE-Qwen2 on all English tasks
- ColSmolVLM is more memory efficient than ColQwen2 💗
Find the model here huggingface.co/vidore/colsm...
Are there account on here that talk about using AI for vulnerability research, offensive security etc? Reading a few papers like this one, give the impression that security researchers may not be utilizing to the full extent multi-turn capabilities or code rag.
arxiv.org/abs/2311.161...
arxiv.org/abs/2311.161...
Understanding the Effectiveness of Large Language Models in Detecting Security Vulnerabilities
While automated vulnerability detection techniques have made promising progress in detecting security vulnerabilities, their scalability and applicability remain challenging. The remarkable performanc...
arxiv.org
November 27, 2024 at 9:01 AM
Are there account on here that talk about using AI for vulnerability research, offensive security etc? Reading a few papers like this one, give the impression that security researchers may not be utilizing to the full extent multi-turn capabilities or code rag.
arxiv.org/abs/2311.161...
arxiv.org/abs/2311.161...
Reposted by Saahil Ognawala
Reposted by Saahil Ognawala
Jina-CLIP-v2: a 0.9B multilingual multimodal embedding model that supports 89 languages, 512x512 image resolution, 8192 token-length, and Matryoshka representations down to 64-dim for both images and text. jina.ai/news/jina-cl... With of course strong performance on retrieval & classification tasks.
Jina CLIP v2: Multilingual Multimodal Embeddings for Text and Images
Jina-CLIP v2, a 0.9B multimodal embedding model with multilingual support of 89 languages, high image resolution at 512x512, and Matryoshka representations.
jina.ai
November 26, 2024 at 8:56 AM
Jina-CLIP-v2: a 0.9B multilingual multimodal embedding model that supports 89 languages, 512x512 image resolution, 8192 token-length, and Matryoshka representations down to 64-dim for both images and text. jina.ai/news/jina-cl... With of course strong performance on retrieval & classification tasks.
Reposted by Saahil Ognawala
✨ Jina AI just released Jina-CLIP-v2: A multimodal (images and texts) & multilingual embedding model. Details in 🧵
Model: huggingface.co/jinaai/jina-...
📈 Jina-CLIP-v2 outperforms Jina-CLIP-v1 (by 3% on text-image and text-text tasks)
🧵
Model: huggingface.co/jinaai/jina-...
📈 Jina-CLIP-v2 outperforms Jina-CLIP-v1 (by 3% on text-image and text-text tasks)
🧵
November 25, 2024 at 9:43 AM
✨ Jina AI just released Jina-CLIP-v2: A multimodal (images and texts) & multilingual embedding model. Details in 🧵
Model: huggingface.co/jinaai/jina-...
📈 Jina-CLIP-v2 outperforms Jina-CLIP-v1 (by 3% on text-image and text-text tasks)
🧵
Model: huggingface.co/jinaai/jina-...
📈 Jina-CLIP-v2 outperforms Jina-CLIP-v1 (by 3% on text-image and text-text tasks)
🧵
Reposted by Saahil Ognawala
November 25, 2024 at 9:29 AM