Version 2 feels quite different from the original, with a heavy focus on reducing repetition across conversations and improving instruction following.

Has a pretty unique writing style and sense of creativity (IMO). Pays the price with intermittent brain farts though.

SillyTavern Settings

Recommended Roleplay Format

> Actions: In plaintext

> Dialogue: "In quotes"

> Thoughts: *In asterisks*

Suggested Samplers

> Temp: 0.5-0.6

> MinP: 0.1

> TopP: 0.95

> Dry: 0.8, 1.75, 4

Instruct

Mistral v7 Tekken

Quantizations

GGUF

> Static (mradermacher)

> iMatrix (mradermacher)

EXL3

> 3bpw

> 3.5bpw

> 4bpw

> 5bpw

> 6bpw

Training Process

Training process: SFT > DPO > KTO

SFT with RP/ERP, Stories and in character assistant data.

DPO focused on reducing repetition, misgendered characters and slop.

KTO focused on further reducing repetition and slop.

Not optimized for cost / performance efficiency, YMMV.

SFT 1*H100

# ====================
# MODEL CONFIGURATION
# ====================
base_model: ConicCat/Mistral-Small-3.2-AntiRep-24B
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer
chat_template: mistral_v7_tekken
[object Object]
# ====================
# DATASET CONFIGURATION
# ====================
datasets:
  - path: ./dataset.jsonl
    type: chat_template
    split: train
    chat_template_strategy: tokenizer
    field_messages: messages
    message_property_mappings:
      role: role
      content: content
    roles:
      user: ["user"]
      assistant: ["assistant"]
      system: ["system"]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object][object Object]

async0x42/MS3.2-PaintedFantasy-v2-24B-exl3_4.5bpw

PAINTED FANTASY v2

Overview