huihui_ai/
exaone3.5-abliterated:2.4b-fp16

1,680 9 months ago

EXAONE 3.5 is a collection of instruction-tuned bilingual (English and Korean) generative models ranging from 2.4B to 32B parameters, developed and released by LG AI Research.

2.4b 7.8b 32b

9 months ago

6ad4bd255e23 · 5.3GB ·

model
exaone
·
2.67B
·
F16
template
{{- range $i, $_ := .Messages }} {{- $last := eq (len (slice $.Messages $i)) 1 -}} {{ if eq .Role "s
system
You are EXAONE model from LG AI Research, a helpful assistant.
license
EXAONE AI Model License Agreement 1.1 - NC This License Agreement (“Agreement”) is entered into
params
{ "repeat_penalty": 1, "stop": [ "[|endofturn|]" ], "temperature": 1 }

Readme

This is an uncensored version of LGAI-EXAONE/exaone-35 created with abliteration (see remove-refusals-with-transformers to know more about it).
This is a crude, proof-of-concept implementation to remove refusals from an LLM model without using TransformerLens.

References

HuggingFace