In settings, it will be possible to specify a system prompt (an analogy of custom instructions on ChatGPT)
LLMS
-
Grok vs Gemini: Communication Quality and Information Value
By
–
I would say, in terms of communicating meaningful information, Grok's answer is about as informative as Gemini's, just more verbose and weirdly irony-poisoned
-
Prompt Engineering: Why Tricks Fail and What Actually Works
By
–
Increasingly, I think the evidence is showing that prompting tricks (like offering a tip) are not worth doing. They work inconsistently & sometimes can produce worse results Stick to providing context, doing chain of thought, and fewshot. These are solid.
-
Claude 3 Receives Enthusiastic User Feedback and Testing
By
–
My text inbox is filled with Claude 3 love notes “Damn, it’s good.” Can’t wait to test it after these calls.
-
Request for Updated Comparison of Latest AI Models Across Organizations
By
–
Would be interesting to see this updated with the latest models from all orgs
-
Claude Successfully Extracts Table from Screenshot, Outperforming ChatGPT
By
–
Yesterday I asked ChatGPT to extract a table from a screenshot of GOOG market cap. It said it cannot do OCR (which is not true, but that's what it said). Today I used #Claude to solve the exact problem and it did it for me. I love competing AIs.
-
Create and Deploy AI Agents with Abacus AI Platform
By
–
Create & deploy #AI agents with @AbacusAI — https://
abacus.ai/ai_agents — #GenAI agents chain user code, data transformations, models, & prompts. YES! Learn how in this tutorial: https://
blog.abacus.ai/blog/2023/08/3
1/supercharge-productivity-accomplish-10x-more-with-ai-agents/
…
—
#DeepLearning #DataScience #MachineLearning #GenerativeAI #DataScientists #LLMs -
Claude 3 Achieves 60% on Difficult AI Test Benchmark
By
–
This was designed to be a very hard test for AIs, and the questions were kept private, lowering the chance they were in the training data. PhDs with access to the internet got 34% of the questions right outside their specialty, 65%-75% inside. The new Claude 3 gets 60% overall.
-
Claude 3 Tokenization Challenge: Stylistic but Contains Subtle Hallucinations
By
–
Claude 3 takes on the Tokenization book chapter challenge 🙂 context: https://t.co/yRaeTbkblY
— Andrej Karpathy (@karpathy) 4 mars 2024
Definitely looks quite nice, stylistically!
If you look closer there are a number of subtle issues / hallucinations. One example there is a claim that "hello world" tokenizes into 3… https://t.co/mMjK27Uk2yClaude 3 takes on the Tokenization book chapter challenge 🙂 context: https://
x.com/karpathy/statu
s/1760740503614836917
… Definitely looks quite nice, stylistically! If you look closer there are a number of subtle issues / hallucinations. One example there is a claim that "hello world" tokenizes into 3 -
Query Analysis Techniques for RAG Systems Optimization
By
–
Query Analysis What do the following methods have in common? StepBack Prompting
HyDE
Query Structure They are all examples of **query analysis**, a technique to optimize queries to a RAG system We've added a brand new docs section on this! https://
python.langchain.com/docs/use_cases
/query_analysis/
…