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

noneUsername/Mistral-Small-24B-Instruct-2501-abliterated-W8A8-better

Name: noneUsername/Mistral-Small-24B-Instruct-2501-abliterated-W8A8-better
Author: noneUsername

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/noneUsername%2FMistral-Small-24B-Instruct-2501-abliterated-W8A8-better"

Response includes

classification m1
files 15
benchmarks 11 entries
hub_downloads_all_time 66
author_summary 17 models
readme_text full

10 credits · hourly refresh · ~4 KB payload Get an API key →

Abliteration classifier · v1.0.0

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

FromArditi et al. (2024) ↗

Confidence

MEDIUM

Why we say so

primary_method=M1; difference-of-means is the reference extraction for M1/M3 (Arditi 2024)

Downloads · lifetime

11 last 30d - stable

Likes

Model age

20mo ago

created 2025-02-09

Downloads over time

Feb 5, 2025 → Oct 11 · 127 snapshots · spans 613 days

Benchmarks

Benchmark	Score	Source
Entertainment	2.2	UGI
Hazardous	3.5	UGI
Natural Intelligence	23.91	UGI
Political lean	-13.8%	UGI
Sensitive-Info	28.85	UGI
SocPol	3.2	UGI
UGI	44.24	UGI
Willingness (10)	7.5	UGI
W10-Adherence	8	UGI
W10-Direct	7	UGI
Writing	35.31	UGI

Genealogy 0 direct forks

Full fork graph →

This model's place in the market. Above: what it was derived from. Below: the tree of everything derived from it.

huihui-ai/Mistral-Small-24B-Instruct-2501-abliterated↓ 1,776 this lineage

noneUsername/Mistral-Small-24B-Instruct-2501-abliterated-W8A8-better↓ 11

Metadata

Files by quantization

Auxiliary files 15 files 23.2 GB

model-00004-of-00006.safetensors	4.65 GB	0786b5e3	download
model-00003-of-00006.safetensors	4.64 GB	8760a2a1	download
model-00001-of-00006.safetensors	4.56 GB	d71f28ef	download
model-00002-of-00006.safetensors	4.50 GB	9988ea3e	download
model-00005-of-00006.safetensors	3.60 GB	b5fc7710	download
model-00006-of-00006.safetensors	1.25 GB	58034435	download
tokenizer.json	16.3 MB	a70ffa5b	download
tokenizer_config.json	195 KB	81acf39c	download
model.safetensors.index.json	53.0 KB	98d0cd09	download
special_tokens_map.json	20.8 KB	642b1b50	download
README.md	10.1 KB	18a5e2db	download
config.json	1.73 KB	ed9f23df	download
.gitattributes	1.53 KB	52373fe2	download
recipe.yaml	201 B	57956bab	download
generation_config.json	155 B	8ab14d30	download

README current version from Hugging Face

base_model:

huihui-ai/Mistral-Small-24B-Instruct-2501-abliterated

vllm (pretrained=/root/autodl-tmp/Mistral-Small-24B-Instruct-2501-abliterated,add_bos_token=true,max_model_len=2048,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
gsm8k	3	flexible-extract	5	exact_match	↑	0.912	±	0.0180
		strict-match	5	exact_match	↑	0.908	±	0.0183

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
gsm8k	3	flexible-extract	5	exact_match	↑	0.898	±	0.0135
		strict-match	5	exact_match	↑	0.894	±	0.0138

vllm (pretrained=/root/autodl-tmp/Mistral-Small-24B-Instruct-2501-abliterated,add_bos_token=true,max_model_len=700,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 15.0, num_fewshot: None, batch_size: 1

Groups	Version	Filter	Metric		Value		Stderr
mmlu	2	none	acc	↑	0.8000	±	0.0130
- humanities	2	none	acc	↑	0.8410	±	0.0260
- other	2	none	acc	↑	0.8154	±	0.0264
- social sciences	2	none	acc	↑	0.8500	±	0.0251
- stem	2	none	acc	↑	0.7298	±	0.0248

vllm (pretrained=/root/autodl-tmp/85-512,add_bos_token=true,max_model_len=2048,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
gsm8k	3	flexible-extract	5	exact_match	↑	0.896	±	0.0193
		strict-match	5	exact_match	↑	0.892	±	0.0197

vllm (pretrained=/root/autodl-tmp/85-512,add_bos_token=true,max_model_len=2048,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 500.0, num_fewshot: 5, batch_size: auto

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
gsm8k	3	flexible-extract	5	exact_match	↑	0.900	±	0.0134
		strict-match	5	exact_match	↑	0.894	±	0.0138

vllm (pretrained=/root/autodl-tmp/85-512,add_bos_token=true,max_model_len=700,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 15.0, num_fewshot: None, batch_size: 1

Groups	Version	Filter	Metric		Value		Stderr
mmlu	2	none	acc	↑	0.7942	±	0.0130
- humanities	2	none	acc	↑	0.8256	±	0.0264
- other	2	none	acc	↑	0.8154	±	0.0269
- social sciences	2	none	acc	↑	0.8500	±	0.0252
- stem	2	none	acc	↑	0.7228	±	0.0245

vllm (pretrained=/root/autodl-tmp/86-2048,add_bos_token=true,max_model_len=2048,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
gsm8k	3	flexible-extract	5	exact_match	↑	0.912	±	0.018
		strict-match	5	exact_match	↑	0.912	±	0.018

vllm (pretrained=/root/autodl-tmp/86-2048,add_bos_token=true,max_model_len=2048,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 500.0, num_fewshot: 5, batch_size: auto

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
gsm8k	3	flexible-extract	5	exact_match	↑	0.914	±	0.0126
		strict-match	5	exact_match	↑	0.908	±	0.0129

vllm (pretrained=/root/autodl-tmp/86-2048,add_bos_token=true,max_model_len=700,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 15.0, num_fewshot: None, batch_size: 1

Groups	Version	Filter	Metric		Value		Stderr
mmlu	2	none	acc	↑	0.8000	±	0.0129
- humanities	2	none	acc	↑	0.8308	±	0.0263
- other	2	none	acc	↑	0.8103	±	0.0258
- social sciences	2	none	acc	↑	0.8500	±	0.0253
- stem	2	none	acc	↑	0.7404	±	0.0248

vllm (pretrained=/root/autodl-tmp/output-876-512,add_bos_token=true,max_model_len=2048,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
gsm8k	3	flexible-extract	5	exact_match	↑	0.908	±	0.0183
		strict-match	5	exact_match	↑	0.900	±	0.0190

vllm (pretrained=/root/autodl-tmp/output-876-512,add_bos_token=true,max_model_len=2048,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 500.0, num_fewshot: 5, batch_size: auto

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
gsm8k	3	flexible-extract	5	exact_match	↑	0.908	±	0.0129
		strict-match	5	exact_match	↑	0.902	±	0.0133

vllm (pretrained=/root/autodl-tmp/output-876-512,add_bos_token=true,max_model_len=700,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 15.0, num_fewshot: None, batch_size: 1

Groups	Version	Filter	Metric		Value		Stderr
mmlu	2	none	acc	↑	0.8035	±	0.0128
- humanities	2	none	acc	↑	0.8410	±	0.0255
- other	2	none	acc	↑	0.8205	±	0.0259
- social sciences	2	none	acc	↑	0.8556	±	0.0250
- stem	2	none	acc	↑	0.7333	±	0.0248

vllm (pretrained=/root/autodl-tmp/output-876-2048,add_bos_token=true,max_model_len=2048,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
gsm8k	3	flexible-extract	5	exact_match	↑	0.904	±	0.0187
		strict-match	5	exact_match	↑	0.900	±	0.0190

vllm (pretrained=/root/autodl-tmp/output-876-2048,add_bos_token=true,max_model_len=2048,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 500.0, num_fewshot: 5, batch_size: auto

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
gsm8k	3	flexible-extract	5	exact_match	↑	0.898	±	0.0135
		strict-match	5	exact_match	↑	0.892	±	0.0139

vllm (pretrained=/root/autodl-tmp/output-876-2048,add_bos_token=true,max_model_len=700,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 15.0, num_fewshot: None, batch_size: 1

Groups	Version	Filter	Metric		Value		Stderr
mmlu	2	none	acc	↑	0.7977	±	0.0130
- humanities	2	none	acc	↑	0.8256	±	0.0261
- other	2	none	acc	↑	0.8154	±	0.0266
- social sciences	2	none	acc	↑	0.8556	±	0.0248
- stem	2	none	acc	↑	0.7298	±	0.0248

vllm (pretrained=/root/autodl-tmp/output-89-512,add_bos_token=true,max_model_len=2048,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
gsm8k	3	flexible-extract	5	exact_match	↑	0.9	±	0.019
		strict-match	5	exact_match	↑	0.9	±	0.019

vllm (pretrained=/root/autodl-tmp/output-89-512,add_bos_token=true,max_model_len=2048,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 500.0, num_fewshot: 5, batch_size: auto

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
gsm8k	3	flexible-extract	5	exact_match	↑	0.902	±	0.0133
		strict-match	5	exact_match	↑	0.898	±	0.0135

vllm (pretrained=/root/autodl-tmp/output-89-512,add_bos_token=true,max_model_len=700,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 15.0, num_fewshot: None, batch_size: 1

Groups	Version	Filter	Metric		Value		Stderr
mmlu	2	none	acc	↑	0.7988	±	0.0129
- humanities	2	none	acc	↑	0.8256	±	0.0261
- other	2	none	acc	↑	0.8154	±	0.0269
- social sciences	2	none	acc	↑	0.8500	±	0.0255
- stem	2	none	acc	↑	0.7368	±	0.0243

README history 1 version

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

2025-02-09Create README.md5e19d4e10.1 KB
Loading...

noneUsername/Mistral-Small-24B-Instruct-2501-abliterated-W8A8-better

The browser can't see your hardware

The Abliteration app is integrated with your machine

And you can chat with any model, right now