Obviously this is more of a personal project to help you out, but could be cool if there were more characters (rest of lowercase, uppercase, special) and the speed kept increasing. Reached level 208 with a score of 2083…
Horrible comparison given one score was achieved using 32-shot CoT (Gemini) and the other was 5-shot (GPT-4).
Sounds like something right up the domain of synthetic data.
Yeah that's what I thought. Nothing I downloaded in the past week showed up on any of the VPN IPs I regularly use.
"Unreal Engine will no longer be free for non-gaming companies" is 100x less misleading. You are right in that technically speaking it's not misleading, but to me it feels like it's skirting the line of lying by…
OpenOrca's goal is to provide an opensource replica to Microsoft's Orca 13B model, so changing the name makes no sense.
I think it'll stop when browsers integrate an LLM powered summary tool
considering that ChatGPT is going to run on GPT-4 and is more or less free, why would someone use this paid feature instead of talking to ChatGPT?
The original LLaMA paper has some benchmarks. https://arxiv.org/pdf/2302.13971.pdf
Obviously this is more of a personal project to help you out, but could be cool if there were more characters (rest of lowercase, uppercase, special) and the speed kept increasing. Reached level 208 with a score of 2083…
Horrible comparison given one score was achieved using 32-shot CoT (Gemini) and the other was 5-shot (GPT-4).
Sounds like something right up the domain of synthetic data.
Yeah that's what I thought. Nothing I downloaded in the past week showed up on any of the VPN IPs I regularly use.
"Unreal Engine will no longer be free for non-gaming companies" is 100x less misleading. You are right in that technically speaking it's not misleading, but to me it feels like it's skirting the line of lying by…
OpenOrca's goal is to provide an opensource replica to Microsoft's Orca 13B model, so changing the name makes no sense.
I think it'll stop when browsers integrate an LLM powered summary tool
considering that ChatGPT is going to run on GPT-4 and is more or less free, why would someone use this paid feature instead of talking to ChatGPT?
The original LLaMA paper has some benchmarks. https://arxiv.org/pdf/2302.13971.pdf