Some considerations:
It works surprisingly well, not many changes needed besides, of course, the different way @ollama inference works. I just had to add some more info in the prompt for the orchestrator because Llama 3 loves to yap, lol. At one point, it forgot the goal and
LLMS
-
Ollama Integration with Llama 3 Orchestration Works Well
By
–
-
Meta Llama 3 Impresses Despite Smaller Model Size
By
–
Meta Llama 3 is very good, especially for such a small model. We can put in a multi-page prompt like our negotiation simulator (
https://
moreusefulthings.com/student-exerci
ses
…) & it is able to follow the complexity reasonably well. It doesn’t have the “smarts” of GPT-4 class, but impressive nonetheless. -
Maestro-Ollama: Local Llama 3 70B Agent Framework
By
–
Introducing Maestro-Ollama! 🦙
— Pietro Schirano (@skirano) 18 avril 2024
You can now harness the power of the Maestro framework entirely locally using Llama 3 70B via @ollama.
Let that sink in for a second, this is a model that outperforms Claude 3 Sonnet, operating as an agent, completely locally.
What a time! 🔥 pic.twitter.com/aC34Bd6F65Introducing Maestro-Ollama! You can now harness the power of the Maestro framework entirely locally using Llama 3 70B via @ollama
. Let that sink in for a second, this is a model that outperforms Claude 3 Sonnet, operating as an agent, completely locally. What a time! -
Fine-tuning Models Before Release for Better Self-Description
By
–
I agree with Jeremy here: we know that people are going to ask models about themselves here whether or not that actually makes sense to do so, so we should be able to fine-tune the model after initial training but before release to help it provide more useful answers there
-
RLHF Annotation Bias and Model Vocabulary Development
By
–
I don't know that RLHF would bias that kind of thing – my mental model is that annotators are shown two answers to the same prompt and asked which is "best", so if none of the test prompts happened to touch on the concept of a roadside kiosk that vocabulary wouldn't be affected
-
Pre-training and Instruction-Tuning in Large Language Models
By
–
The pre-training stage is where the billions of words of training data come into play – the instruction-tuning / RLHF stage is where human labelers are asked to vote on which generations are "best" – that's the bit that might influence things like "delve" https://
openai.com/research/instr
uction-following
… -
RLHF and Instruction Tuning: Understanding Model Training Mechanisms
By
–
Yeah, what did it get wrong? It fitted my mental model of how the RLHF/instruction tuning stage works pretty closely
-
HyperWrite Launches Fine-Tuned LLaMA 3 Model
By
–
An initial HyperWrite fine-tune of LLaMA 3 is working and we're evaluating it in the HyperWrite platform. Will iterate, then ship to users soon!
-
GPT-3 Ada Model: 2.7 Billion Parameters Breakdown
By
–
I learned today that GPT-3 Ada was a 2.7bn parameter model (Babbage was 6.7, Curie 13 and Davinci 175)
-
Prompt Injection vs Jailbreaking: Clarifying Key Security Concepts
By
–
That document has an incorrect definition of prompt injection: it says "Prompt injection attacks are attempts to circumvent content restrictions to produce particular outputs" – but that's not prompt injection, that's jailbreaking