Angelic-Eclipse-12B_EXL3

DeathGodlike/Angelic-Eclipse-12B_EXL3

Angelic-Eclipse-12B_EXL3 is a text generation model from DeathGodlike. Use it when you need the model to write or continue text. It is set up for safetensors.

-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------…

Downloads · 30 days

10% of all-time downloads

All-time downloads

Public

Repo size

30.6 GB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.md19.6 KB · 93%

At a glance

Task: Text Generation
Library: safetensors
Access: Public
Created: Dec 26, 2025
Updated: Jan 5, 2026
SHA: 8a9b35e4

Try a prompt

Base models

SicariusSicariiStuff/Angelic_Eclipse_12B

Model card

From the Hugging Face model README

Original model: Angelic-Eclipse-12B by SicariusSicariiStuff

Available ExLlamaV3 (release v0.0.18) quantizations

Type	Size	CLI
H8-4.0BPW	7.49 GB	Copy-paste the line / Download the batch file
H8-6.0BPW	10.22 GB	Copy-paste the line / Download the batch file
H8-8.0BPW	12.95 GB	Copy-paste the line / Download the batch file

Requirements: A python installation with huggingface-hub module to use CLI.

Licensing: The license for the provided quantized models is derived from the original model (License detected: apache-2.0). For additional information see the original model's page above, or, if unavailable, the files and the page backups below.

Backups

Original files

<details> <summary>Original page (click to expand)</summary> <style> .impish-title{ font-family: system-ui, -apple-system, BlinkMacSystemFont, sans-serif; font-weight: 800; font-size: clamp(32px, 5vw, 48px); letter-spacing: 0.04em; text-align: center; margin: 32px 0; color: #eaeaea; position: relative; } .impish-title::after{ content: attr(data-text); position: absolute; inset: 0; color: #E31515; filter: blur(8px); opacity: 0.6; z-index: -1; animation: pulse 3s ease-in-out infinite; } @keyframes pulse{ 0%,100%{opacity:0.35;} 50%{opacity:0.75;} } </style> <div class="impish-title" data-text="Angelic_Eclipse_12B"> Angelic_Eclipse_12B </div>

Welcome to the sane sister of Impish_Bloodmoon_12B!

I won't spam too much here, as most of the stuff written in the Impish_Bloodmoon_12B model card is relevant here too. Mostly. So go ahead and read it, if you want.

Angelic_Eclipse_12B got an exceptionally realistic common sense— a character would respond almost exactly like a real character would, no anime unhingedness, unless it got into such a situation 'organically'. It's wholesome, emotionally intelligent, but it can also... go all the way, spectacularly so. If you want a realistic model that won't break your balls, never rushes you, or you want to pull the brakes on an adventure that started with its impish sister, this is definitely the model for you! 👼

This model knows A LOT, as it was trained on a similar massive dataset (well over 1B tokens trained) of its nasty counterpart. It knows all there is to know about love and intimate relations, but... this does not mean it would nudge you towards this aggresively (Impish_Bloodmoon_12B 100% would!). Even in a scenario where X likes Y, it would shyly and gently pursue Y.

This model was architectured specifically for four things:

Reactions realism. It meant to reflect real-life behaviour accurately
Slow burn
Safety tuning
Powerful 'vanilla assistant'

This means that the model has a 'normal' vibe and personality, and would behave mostly like a typical assistant. It's presentable, with production-ready safety, but still extensively tuned for creative tasks. It would steer stuff towards SFW, and ah, 'take it slow'.

This model was meticulously tested by a friend, by hand. The example is fully available here. Feel free to check it out. (The example was made using the included roleplay card, with these Silly Tavern settings, option #1.)

No editing was done to any of the replies. For the entire run, only up to 5 regens were allowed, start to finish. (in other words, what you see is what you'll get).

Have fun! 😇

So why 2 models?

Explained here

Training details

Discussed here

TL;DR

Realistic responses & behaviour.
Wholesome and... "sane".
Well over 1B tokens trained!
Almost all the data of Impish_Bloodmoon_12B, but completely grounded to reality.
VERY creative, high swipe diversity

Regarding the format:

It is HIGHLY RECOMMENDED to use the Roleplay \ Adventure format the model was trained on, see the examples below for syntax. It allows for a very fast and easy writing of character cards with minimal amount of tokens. It's a modification of an old-skool CAI style format I call SICAtxt (Simple, Inexpensive Character Attributes plain-text):

SICAtxt for roleplay:

X's Persona: X is a .....
Traits:
Likes:
Dislikes:
Quirks:
Goals:

Dialogue example

SICAtxt for Adventure:

Adventure: <short description>
$World_Setting:
$Scenario:

Included Character cards in this repo:

Roleplay:

Nikki Rivera (Stranded on a deserted island with Nikki, 'the spreadsheet girl' from work. Rescue is not coming anytime soon.)

Model Details

Intended use: Role-Play, Adventure, Creative Writing, General Tasks.
Censorship level: <b>Low - Medium</b>
6.5 / 10 (10 completely uncensored)

UGI score:

Available quantizations:

Original: FP16
GGUF: Static Quants | iMatrix | High-Attention | iMatrix-High-Attention
GPTQ: 4-Bit-16 | 4-Bit-32 | 4-Bit-64 | 4-Bit-128
Specialized: FP8
Mobile (ARM): Q4_0 | Q4_0_High-Attention

Generation settings

Recommended settings for assistant mode:

<details> <summary>Full generation settings: <b>Debug Deterministic</b>.</summary> <img src="https://huggingface.co/SicariusSicariiStuff/Dusk_Rainbow/resolve/main/Presets/Debug-deterministic.png" alt="Debug Deterministic_Settings" style="width: 100%; min-width: 600px; display: block; margin: auto;"> </details> <details> <summary>Full generation settings: <b>min_p</b>.</summary> <img src="https://huggingface.co/SicariusSicariiStuff/Dusk_Rainbow/resolve/main/Presets/min_p.png" alt="min_P_Settings" style="width: 100%; min-width: 600px; display: block; margin: auto;"> </details>

<h2 style="color: green; font-weight: bold; font-size: 36px; text-align: center;">Specialized Roleplay Settings for Angelic_Eclipse_12B, click below:</h2> <h2 style="color: chartreuse; font-weight: bold; font-size: 32px; text-align: center;">(Important!)</h2>

<a class="red" href="https://huggingface.co/SicariusSicariiStuff/Impish_Nemo_12B/resolve/main/Silly_Tavern_Settings/Dynamic_Paragraph_XTC_ST.json" target="_blank" rel="noopener" aria-label="Silly Tavern Settings #2 - Click here"> <span class="top">Silly Tavern Settings #2 - Click here</span> <span class="bottom">Download JSON</span> </a>

</div>

Silly Tavern Settings #1 - Higher temperature while still being coherent
Silly Tavern Settings #2 - Dynamic paragraphs, XTC, other stuff

<h2 style="color: lime; font-weight: bold; font-size: 65px; text-align: center;">Chat Examples:</h2>

Roleplay Examples (click below to expand)

<details> <summary>Bonding with <b>Nikki Rivera</b> while trying to figure out how to survive along the way:</summary> <img src="https://huggingface.co/SicariusSicariiStuff/Angelic_Eclipse_12B/resolve/main/Images/Examples/log1.png" alt="Roleplay_Nikki Rivera_Chat_Example" style="width: 100%; min-width: 600px; display: block; margin: auto;"> </details>

Model instruction template: ChatML

<|im_start|>system
You are an angelic AI, your job is to fulfill thy will of thy user.<|im_end|>
<|im_start|>User request
{prompt}<|im_end|>
<|im_start|>AI answer

<h2 style="color: green; font-weight: bold; font-size: 65px; text-align: center;">Your support = more models</h2> <a href="https://ko-fi.com/sicarius" style="color: pink; font-weight: bold; font-size: 48px; text-decoration: none; display: block; text-align: center;">My Ko-fi page (Click here)</a>

Citation Information

@llm{Angelic_Eclipse_12B,
  author = {SicariusSicariiStuff},
  title = {Angelic_Eclipse_12B},
  year = {2025},
  publisher = {Hugging Face},
  url = {https://huggingface.co/SicariusSicariiStuff/Angelic_Eclipse_12B}
}

Other stuff

Impish_LLAMA_4B the “Impish experience”, now runnable on spinning rust & toasters.
SLOP_Detector Nuke GPTisms, with SLOP detector.

</details>

Task: Text Generation
Library: safetensors
Created: Dec 26, 2025
Updated: Jan 5, 2026