Minimal details of the training data in the model card aside from a list of 23 languages (10 as a focus, 13 additional) that were covered https://
huggingface.co/CohereForAI/c4
ai-command-r-plus
…
@simonw
-
Cohere Command R Plus Model Card Details Released
By
–
-
New 128K Token Model Available with Commercial API Pricing
By
–
128,000 token length, weights available under a non-commercial license or a commercial API, fine-tuned for both tool use and citation-providing RAG $3/M input, $15/M output – same price as Claude 3 Sonnet
-
Prompt Engineering Techniques for Haiku Model Optimization
By
–
The two tricks I'm trying with Haiku that seem to work incredibly well are feeding it a small targeted sequence of fake "user"/"assistant" examples in the messages array and finishing with a prefilled "assistant" message to point it in the right direction
-
Open weights embedding models gaining traction over closed APIs
By
–
I find it hard to get excited about closed API-only embedding models these days, because open weights ones (eg Nomic 1.5) are getting REALLY good I'd rather use on open weights model available via an API, safe in the knowledge that if the API ever goes away I can run it myself
-
Quick Screenshot Tool with Annotations for Team Collaboration
By
–
I use it every day – it's so handy to be able to snap a quick shot of an area of a screen, optionally add some big pink arrows and then drag it straight to Slack or Google Docs or a GitHub Issue
-
Using GitHub Issues as Documentation and Blog Post Source
By
–
I like GitHub issues for this because they're free, reliable, accessible via an API and easy to search later on I often turn my issues into blog posts too – since the comments are already markdown it's not a huge lift to turn them into posts for https://
til.simonwillison.net -
RAG without vectors: LLM keywords search FTS alternative
By
–
Right: you can implement RAG without any vectors at all – get an LLM to spit out some keywords search phrases to run, then run those against a regular FTS search engine (or even with a bunch of SQL LIKE queries) and paste the results back into the context
-
GitLab Data Deletion: Cultural Failing in User Trust
By
–
Normally I'd agree – companies make mistakes. But this was such a big one! For me it speaks to a cultural failing – I would hope that an organization like GitLab would have "don't delete data that our users have trusted to us" as a key principle
-
GitLab Reverses Data Deletion Policy Decision
By
–
Here's that story https://
theregister.com/2022/08/05/git
lab_reverses_deletion_policy/
… -
Creating LLM Embeddings Plugins: Documentation and Examples
By
–
There's documentation on creating a new embeddings plugin here https://
llm.datasette.io/en/stable/embe
ddings/writing-plugins.html
… They are pretty simple though, so best approach is to look at one of the existing embeddings plugins and copy from that – here's one I wrote the other day https://
github.com/simonw/llm-nom
ic-api-embed/blob/main/llm_nomic_api_embed.py
…