Weeknotes: Llama 3, AI for Data Journalism, llm-evals and datasette-secrets
DATA
-
The Three Essential Elements of High-Quality Datasets
By
–
The 3 key elements of a good dataset: 1. quality
2. diversity
3. quantity You can only easily measure the last one but the performance is a sensitive function of all three. Super interesting topic ty for #longread :)! -
Data contamination detection script for training datasets
By
–
Just provide a script that will produce a number from 0 to 1 to be run on your training data on how contaminated it is. Everyone runs it and reports it as the first column of results.
-
Wayfair Improves Model Accuracy Using Programmatic Data Development
By
–
Wayfair improved model accuracy by 20+ points – want to learn how? Join to see how Wayfair and Snorkel used programmatic data development and foundation models to deliver a production model which helps customers search 40 million products. https://
buff.ly/4aXqt9w -
Training costs for 8B parameter models relatively affordable
By
–
Couple hundred. Pretty cheap b/c only 16K on 8B
-
Explainable AI: Challenges Across Telecommunications and Research
By
–
Looking to gain a deeper understanding of Explainable AI? Experts from Eckerson Group, Orange and Georgia Tech dive into the challenges of AI explainability from telecommunications networks to academic research. https://
buff.ly/3vNcwMr -
LangSmith Webhooks with Modal for Custom Automation
By
–
LangSmith Webhooks w/ Modal LangSmith automations help you build a data flywheel by taking actions on data. By default you can move data to datasets or annotation queue With @modal_labs you can set up a webhook to perform arbitrary actions https://
docs.smith.langchain.com/monitoring/faq
/webhooks#example-with-modal
… -
AI Agents Vertical Growing Fast with Dynamic RAG Support
By
–
Our AI Agents vertical is our fastest-growing vertical. It can help you crank out multiple bots to perform different tasks of varying complexity. Some of our customers even want to make 700+ agents! You will need support for:
•Dynamic RAG, information retrieval, and vector -
Data Congruency Critical for Fine-Tuning LLMs in Products
By
–
When fine-tuning LLMs for products, data congruency is so important.
-
Context Window Limitations Challenge RAG Implementation at Scale
By
–
Even 100,000 tokens (or a million tokens) isn't enough for most of the things I'd want to use RAG for – I tried my blog's full archive with Gemini Pro 1.5 and it didn't fit into their million token window