rm is the reward model… not the same as the lm model. I tried the lm, wasn’t impressed. Gpt-3.5 did better for summarizing quotes. It was good, but I honestly think open hermes and or synthia 1.3b do better

Yes I understand all that

Auto regressive is like arima In time series forecasting

Then rnn came along

Then sequence to sequence

They all have the last prediction is used as input for the next prediction in common

Hence auto regressive

I had to read that a few times.

Auto-Regressive is like forecasting, it’s iterative.

LLM reliability is this vague concept of trying to get to the right answer.

Hence tree of thoughts as a way to ‘plan’ to that vague concept of the right answer.

Circumvents the univariate next token prediction limitation with parallel planning.

https://twitter.com/ylecun/status/1728126868342145481?ref_src=twsrc%5Egoogle%7Ctwcamp%5Eserp%7Ctwgr%5Etweet

How would a non compete work in this agreement

I was going to try to knowledge distill but they modified their tokenizer.

Either way neo has a 125M model, so a 248M model is x2 that. I imagine this could be useful for shorter context tasks. Idk, or to continue training for very tight uses cases

I came across it while looking for tiny mistral config jsons to replicate⁸

https://preview.redd.it/l9l7a39u3a1c1.jpeg?width=720&format=pjpg&auto=webp&s=80589cb6fbb2268b0d8af65b4ec27647185b4780

I imagine creating an app, putting it on everyone’s cell phone, and using a fraction of the power, you can build an llm easily that would surpass any single data center.