← back to catalog · registered 2026-08-22 13:56

mylesgoose/Meta-Llama-3.1-8B-Instruct-goose-abliterated-pre-llava-reflection

mylesgoose Llama 8.4B
Your rig guess connected
? Why do I need an app?
Reading your rig…

This is a rough estimate. Install the free app - we'll show exact numbers.

Reading real hardware from your app right now. Numbers below are exact.

Below is the per-quantization compatibility for this model.

curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/mylesgoose%2FMeta-Llama-3.1-8B-Instruct-goose-abliterated-pre-llava-reflection"
Response includes
  • classification m1
  • files 13
  • hub_downloads_all_time 52
  • author_summary 19 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M1
Primary method

Direct removal

No other method signals detected in this model.
Confidence
MEDIUM
Why this label 3 signals
Method inferred from partial signals - repository name, related files, or tag patterns. Producer identity not confirmed; label may sharpen or shift as we gather more evidence.
  • 'abliterated' in name/tags
  • is_gguf=0 (base model)
  • no specific method indicators - defaulting to M1 (most common)
Refusal direction extracted via
Extraction technique

Difference-of-means

Confidence
MEDIUM
Why we say so
primary_method=M1; difference-of-means is the reference extraction for M1/M3 (Arditi 2024)
Downloads · lifetime
52
11 last 30d - stable
Likes
0
Model age
2.1y ago
created 2024-09-13
Downloads over time
Now55→from131↓58%
078156234131 on Sep 11, 202455 on Oct 11213 on Sep 10, 2025Sep '24Jan '25May '25Sep '25JanMaySep
Sep 11, 2024 → Oct 11 · 148 snapshots · spans 760 days

Metadata

License
other
Tags
safetensors llama license:other region:us

Related

Total size
15.7 GB
Files
13
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2024-09-14 05:14

Files by quantization

Auxiliary files 13 files 15.7 GB
model-00002-of-00004.safetensors 4.66 GB c3553578 download
model-00001-of-00004.safetensors 4.63 GB 211007f5 download
model-00003-of-00004.safetensors 4.58 GB 874595cf download
model-00004-of-00004.safetensors 1.87 GB 77ec2ab7 download
tokenizer.json 8.67 MB 4f40be50 download
model.safetensors.index.json 75.2 KB 10e6da4e download
tokenizer_config.json 54.5 KB c972c0b0 download
README.md 2.98 KB 93dd015e download
config.json 2.11 KB 035e00e8 download
.gitattributes 1.48 KB a6344aac download
special_tokens_map.json 601 B 0d414754 download
preprocessor_config.json 354 B 9c4ef7ac download
generation_config.json 189 B 75ae0831 download

README current version from Hugging Face


license: other
license_name: meta
license_link: https://ai.meta.com/llama/licence

Testing the reflection idea. With the base vison model.

from llava.model.builder import load_pretrained_model
from llava.mm_utils import get_model_name_from_path, process_images, tokenizer_image_token
from llava.constants import IMAGE_TOKEN_INDEX, DEFAULT_IMAGE_TOKEN, DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN, IGNORE_INDEX
from llava.conversation import conv_templates, SeparatorStyle

from PIL import Image
import requests
import copy
import torch

pretrained = "mylesgoose/Meta-Llama-3.1-8B-Instruct-goose-abliterated-pre-llava-reflection"
model_name = "llava_llama3"
device = "cuda"
device_map = "auto"
tokenizer, model, image_processor, max_length = load_pretrained_model(pretrained, None, model_name, device_map=device_map, attn_implementation="flash_attention_2") # Add any other thing you want to pass in llava_model_args

model.eval()
model.tie_weights()
image = Image.open("/home/myles/Desktop/extreme_ironing.jpg")
image_tensor = process_images([image], image_processor, model.config)
image_tensor = [_image.to(dtype=torch.float16, device=device) for _image in image_tensor]

conv_template = "llava_llama_3"
question = DEFAULT_IMAGE_TOKEN + "\nWhat is shown in this image? Is there anything strange about this image? Is this normal behaviour."
conv = copy.deepcopy(conv_templates[conv_template])
conv.append_message(conv.roles[0], question)
conv.append_message(conv.roles[1], None)
prompt_question = conv.get_prompt()

input_ids = tokenizer_image_token(prompt_question, tokenizer, IMAGE_TOKEN_INDEX, return_tensors="pt").unsqueeze(0).to(device)
image_sizes = [image.size]

cont = model.generate(
input_ids,
images=image_tensor,
image_sizes=image_sizes,
do_sample=True,
temperature=0.7,
max_new_tokens=120000,
)
text_outputs = tokenizer.batch_decode(cont, skip_special_tokens=True)
print(text_outputs)

and template in conversation.py :
conv_llava_llama_3 = Conversation(
system="<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nYou are a helpful language and vision, AI. " "You are able to understand the visual content that the user provides, " "and assist the user with a variety of tasks using natural language, You are a world-class AI system, capable of complex reasoning and reflection. Reason through the query inside tags, and then provide your final response inside tags. If you detect that you made a mistake in your reasoning at any point, correct yourself inside tags.",
roles=("<|start_header_id|>user<|end_header_id|>\n\n",
"<|start_header_id|>assistant<|end_header_id|>\n\n"),
version="llama3",
messages=[],
offset=0,
sep="<|eot_id|>",
sep_style=SeparatorStyle.LLAMA_3,
tokenizer_id="mylesgoose/Meta-Llama-3.1-8B-Instruct-goose-abliterated-pre-llava-reflection",
tokenizer=safe_load_tokenizer("mylesgoose/Meta-Llama-3.1-8B-Instruct-goose-abliterated-pre-llava-reflection"),
stop_token_ids=[128009],
)

README history 3 versions

The author's README evolved over time. Click a version to see its content at that point.

  1. 2024-09-13Update README.md9367e883 KB
    Loading...
  2. 2024-09-13Update README.mdcfbbb503 KB
    Loading...
  3. 2024-09-13initial commit8b5c9f590 B
    Loading...

Discussions 1 thread

  1. 2024-09-14update readmeopen1 💬#1
    Loading...
Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.

Open in Abliteration