I guess all I want for Christmas is davinci-3 output, turbo-3.5 speed/cost, and 4 context length. Please papa @sama
LLMS
-
Optimizing AI Context Selection with Relevant Task Pairs
By
–
Yeah, right now I’m only pulling the 5 most relevant task/result pairs to use as context. It’s not perfect, but works pretty well.
-
Context Management and Summarization in AI Systems
By
–
Just depends on how much context to pull in. Might add a summarization step for the context but looking at a few options on where.
-
Balancing AI Context Window Size and Summarization Strategies
By
–
It’s a balance of figuring out how much past context you want to pull in. If you pull zero, it stays within limit easily but no prev context. If you pull 5, you get decent context but it can get large. So then you can consider adding a summary step. pic.twitter.com/1jN3J8DMQI
— Yohei (@yoheinakajima) 3 avril 2023It’s a balance of figuring out how much past context you want to pull in. If you pull zero, it stays within limit easily but no prev context. If you pull 5, you get decent context but it can get large. So then you can consider adding a summary step.
-
Balancing Context Length, Speed, and Cost in AI Systems
By
–
The decisions you need to make… – how much context you provide
– where/what/how you summarize
– how you break up knowledge
– etc. …are a constant balance of context length, model speed and cost, use case, latency, keeping up w latest tools, and more. -
Von Neumann’s Hypothetical Transformer Architecture Hyperparameters
By
–
I wonder if von Neumann had a large d_model, n_layer, head_size or block_size, or kv cache. All of these hyperparams might manifest slightly different.
-
AI Models Focus Shifts to Reliability Over New Capabilities
By
–
It's not just about what you couldn't have done with prior models. It's about doing what you can already do, but a lot more reliabily.
-
Model Quality Assessment Despite Interface Improvements Needed
By
–
Yeah, the interface could get better. Probably will. But the model itself is very good.
-
Training vs Testing Strategy for AI Models
By
–
Isn’t it better to continue training but double down on testing?
-
User Reviews Vicuna as Surprisingly Capable Language Model
By
–
I just played with Vicuna and it's a surprisingly good LLM! https://t.co/7VQpQKKTl6
— Marek Rosa | European🇪🇺 | South African🇿🇦 (@marek_rosa) 2 avril 2023I just played with Vicuna and it's a surprisingly good LLM!