131072 Context
$0.390 / 1M input tokens
$0.390 / 1M output tokens
Demo
API
Model Configuration
Response format
System Prompt
max_tokens
temperature
top_p
min_p
top_k
presence_penalty
frequency_penalty
repetition_penalty
README

She's back!

Stheno's Sister Model, designed to impress.

1- Same Dataset used as Stheno v3.2 -> See notes there. 2- LoRA Fine-Tune -> FFT is simply too expensive. 3- Trained over 8x H100 SXMs and then some more afterwards.

Testing Notes

1- Better prompt adherence. 2- Better anatomy / spatial awareness. 3- Adapts much better to unique and custom formatting / reply formats. 4- Very creative, lots of unique swipes. 5- Is not restrictive during roleplays. 6- Feels like a big brained version of Stheno.

Likely due to it being a 70B model instead of 8B. Similar vibes comparing back in llama 2, where 70B models were simply much more 'aware' in the subtler areas and contexts a smaller model like a 7B or 13B simply were not able to handle.


Recommended Sampler Settings :

1Temperature - 1.17 2min_p - 0.075 3Repetition Penalty - 1.10

SillyTavern Instruct Settings : Context Template: Llama-3-Instruct-Names Instruct Presets: Euryale-v2.1-Llama-3-Instruct


As per usual, support me here:

Ko-fi: https://ko-fi.com/sao10k

1Art by wada_kazu / わだかず (pixiv page private?)