exl2 quant:

An (as of right now) experimental/testing merge to try and increase Stew's capabilities, but with some slight alterations in models used, and this one actually seems to show a bit more promise than v3 with the brief tests done so far.

trust-remote-code must be turned on for this version still due to the base model being Capybara, but I'll look into fixing this later if it performs comparably to v2 or better during further testing.

Temperature @ 0.95

Min-P @ 0.1

Smoothing Factor @ 0.3

DRY Multiplier (plus standard DRY settings) @ 0.8

Skip Special Tokens @ On

Everything else @ Off

SYSTEM: {system_prompt}<|end|>
USER: {prompt}<|end|>
ASSISTANT: {output}<|end|>

The following models were included in the merge: