It is pretty amazing that the first 1,000 words recovered from the charred scrolls at Herculaneum taught us two new things: where Plato was buried and that he was sold as a slave by the Spartans. Usually, most data you get is boring, a promising start! https://
cnr.it/it/nota-stampa
/n-12655/lo-sguardo-tecnologico-legge-i-papiri-carbonizzati
…
RESEARCH
-
AI Technology Deciphers Ancient Herculaneum Scrolls
By
–
-
Professional Organizations Setting AI Benchmarks Standards
By
–
This would also be a good thing for professional organizations to do. AMA benchmarks, New York Bar Association benchmarks, Modern Language Association benchmarks, American Psychological Association benchmarks…
-
Snowflake Arctic: Massive Open-Source Model for SQL and Coding
By
–
Snowflake Arctic is a new open-source language model from @SnowflakeDB
. – It's absolutely massive (480B)
– It's very good at SQL and other coding tasks
– It cost just under $2M to train (less than half of Llama) Try it out: https://
arctic.streamlit.app -
Apple Publishes Major AI Research Paper
By
–
Apple dropping absolute FACTS! Give them kudos: https://
huggingface.co/papers/2404.14
619
… -
SciSpace Copilot provides real-time answers for online articles.
By
–
7. SciSpace Copilot
— God of Prompt (@godofprompt) 24 avril 2024
SciSpace Copilot gives you real-time answers to articles, no matter where you read them online.https://t.co/gIxDtx6lTz pic.twitter.com/pLOAgmiOcR7. SciSpace Copilot SciSpace Copilot gives you real-time answers to articles, no matter where you read them online. https://
typeset.io/copilot-chrome
-plugin
… -
PhotoRoom Builds Custom Image Model Instead of Relying Closed AI
By
–
Interesting in-depth read from @BenTheEgg about how PhotoRoom trained their own image model from scratch Another proof point that you should build your own ML capabilities rather than rely on closed AI
-
LLMs analyzing text at runtime capabilities
By
–
For training yes, but I'm talking about LLMs analyzing text at runtime
-
Leaderboards Insufficient Generic Testing Standards AI
By
–
Leaderboards are fine but not a good generic testing standard. And also likely game-able
-
AI Needs Independent Testing Standards Like Consumer Reports
By
–
We need a Consumer Reports or Underwriters Lab for AI testing. All the public benchmarks are game-able and mostly not useful measures of things LLMs do. We need secret test batteries for subject areas (coding, reasoning, human conversation, writing) and secret red team tests, too
-
OpenELM Training Data Sources Documentation
By
–
This is the best documentation I could find of the OpenELM training data – it looks like the bulk of it comes from RefinedWeb, RedPajama, The Pile and Dolma https://
github.com/apple/corenet/
blob/main/projects/openelm/README-pretraining.md
…