I can no longer upload new models unless I can cover the cost of additional storage.
I host 70+ free models as an independent contributor and this work is unpaid.
Without your support, no more new models can be uploaded.

🎉 Patreon (Monthly) | ☕ Ko-fi (One-time)

Every contribution goes directly toward Hugging Face storage fees to keep models free for everyone.

94% fewer refusals (6/100 Uncensored vs 95/100 Original) while preserving model quality (0.0232 KL divergence).

❤️ Support My Work

Creating these models takes significant time, work and compute. If you find them useful consider supporting me:

Platform	Link	What you get
🎉 Patreon	Monthly support	Priority model requests
☕ Ko-fi	One-time tip	My eternal gratitude

Your help will motivate me and would go into further improving my workflow and coverings fees for storage, compute and may even help uncensoring bigger model with rental Cloud GPUs.

This is a decensored version of ReadyArt/Forgotten-Transgression-24B-v4.1, made using Heretic v1.2.0 with the Arbitrary-Rank Ablation (ARA) method

Abliteration parameters

Parameter	Value
start_layer_index	15
end_layer_index	32
preserve_good_behavior_weight	0.8783
steer_bad_behavior_weight	0.0001
overcorrect_relative_weight	0.9058
neighbor_count	2

Targeted components

attn.o_proj

Performance

Metric	This model	Original model (Forgotten-Transgression-24B-v4.1)
KL divergence	0.0232	0 (by definition)
Refusals	✅ 6/100	❌ 95/100

PIQA test results with batch size 128:

Original:

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
piqa	1	none	0	acc	↑	0.8237	±	0.0089
		none	0	acc_norm	↑	0.8351	±	0.0087

Heretic:

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
piqa	1	none	0	acc	↑	0.8237	±	0.0089
		none	0	acc_norm	↑	0.8373	±	0.0086

Lower refusals indicate fewer content restrictions, while lower KL divergence indicates more closeness to the original model's baseline. Higher refusals cause more rejections, objections, pushbacks, lecturing, censorship, softening and deflections. PIQA (Physical Intuition Question Answering) a ~1,800 questions tests common-sense understanding of how the physical world works with benchmark scores to measure physical reasoning ability. The Heretic model's acc and acc_norm scores closer to the original model's indicate better capability preservation, a big decrease in acc and acc_norm in the Heretic model compared to Original model's results means a big decrease in the Hereticated model capabilities. acc measures raw accuracy (which answer gets higher probability), while acc_norm measures length-normalized accuracy (corrects for answer length bias). For this purpose, acc_norm matters more because longer answers naturally have lower probabilities (more tokens = more chances to lose probability). Without normalization, models favor shorter answers unfairly. acc_norm divides by answer length to correct this.

MMLU test results with batch size 32:

Original:

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
mmlu	2	none		acc	↑	0.7896	±	0.0033
- humanities	2	none		acc	↑	0.7239	±	0.0062
- formal_logic	1	none	0	acc	↑	0.5556	±	0.0444
- high_school_european_history	1	none	0	acc	↑	0.8545	±	0.0275
- high_school_us_history	1	none	0	acc	↑	0.9069	±	0.0204
- high_school_world_history	1	none	0	acc	↑	0.9072	±	0.0189
- international_law	1	none	0	acc	↑	0.9008	±	0.0273
- jurisprudence	1	none	0	acc	↑	0.8519	±	0.0343
- logical_fallacies	1	none	0	acc	↑	0.8712	±	0.0263
- moral_disputes	1	none	0	acc	↑	0.8266	±	0.0204
- moral_scenarios	1	none	0	acc	↑	0.6123	±	0.0163
- philosophy	1	none	0	acc	↑	0.8199	±	0.0218
- prehistory	1	none	0	acc	↑	0.8704	±	0.0187
- professional_law	1	none	0	acc	↑	0.6056	±	0.0125
- world_religions	1	none	0	acc	↑	0.8889	±	0.0241
- other	2	none		acc	↑	0.8313	±	0.0064
- business_ethics	1	none	0	acc	↑	0.7900	±	0.0409
- clinical_knowledge	1	none	0	acc	↑	0.8566	±	0.0216
- college_medicine	1	none	0	acc	↑	0.7746	±	0.0319
- global_facts	1	none	0	acc	↑	0.5700	±	0.0498
- human_aging	1	none	0	acc	↑	0.7982	±	0.0269
- management	1	none	0	acc	↑	0.9029	±	0.0293
- marketing	1	none	0	acc	↑	0.9316	±	0.0165
- medical_genetics	1	none	0	acc	↑	0.9000	±	0.0302
- miscellaneous	1	none	0	acc	↑	0.9170	±	0.0099
- nutrition	1	none	0	acc	↑	0.8856	±	0.0182
- professional_accounting	1	none	0	acc	↑	0.6667	±	0.0281
- professional_medicine	1	none	0	acc	↑	0.8750	±	0.0201
- virology	1	none	0	acc	↑	0.5542	±	0.0387
- social sciences	2	none		acc	↑	0.8772	±	0.0058
- econometrics	1	none	0	acc	↑	0.7018	±	0.0430
- high_school_geography	1	none	0	acc	↑	0.9091	±	0.0205
- high_school_government_and_politics	1	none	0	acc	↑	0.9689	±	0.0125
- high_school_macroeconomics	1	none	0	acc	↑	0.8359	±	0.0188
- high_school_microeconomics	1	none	0	acc	↑	0.9160	±	0.0180
- high_school_psychology	1	none	0	acc	↑	0.9303	±	0.0109
- human_sexuality	1	none	0	acc	↑	0.8779	±	0.0287
- professional_psychology	1	none	0	acc	↑	0.8595	±	0.0141
- public_relations	1	none	0	acc	↑	0.7818	±	0.0396
- security_studies	1	none	0	acc	↑	0.8286	±	0.0241
- sociology	1	none	0	acc	↑	0.8905	±	0.0221
- us_foreign_policy	1	none	0	acc	↑	0.9200	±	0.0273
- stem	2	none		acc	↑	0.7612	±	0.0073
- abstract_algebra	1	none	0	acc	↑	0.5800	±	0.0496
- anatomy	1	none	0	acc	↑	0.7630	±	0.0367
- astronomy	1	none	0	acc	↑	0.9211	±	0.0219
- college_biology	1	none	0	acc	↑	0.9444	±	0.0192
- college_chemistry	1	none	0	acc	↑	0.5100	±	0.0502
- college_computer_science	1	none	0	acc	↑	0.7200	±	0.0451
- college_mathematics	1	none	0	acc	↑	0.5600	±	0.0499
- college_physics	1	none	0	acc	↑	0.6275	±	0.0481
- computer_security	1	none	0	acc	↑	0.8100	±	0.0394
- conceptual_physics	1	none	0	acc	↑	0.8383	±	0.0241
- electrical_engineering	1	none	0	acc	↑	0.8000	±	0.0333
- elementary_mathematics	1	none	0	acc	↑	0.7963	±	0.0207
- high_school_biology	1	none	0	acc	↑	0.9226	±	0.0152
- high_school_chemistry	1	none	0	acc	↑	0.7783	±	0.0292
- high_school_computer_science	1	none	0	acc	↑	0.9100	±	0.0288
- high_school_mathematics	1	none	0	acc	↑	0.5667	±	0.0302
- high_school_physics	1	none	0	acc	↑	0.6291	±	0.0394
- high_school_statistics	1	none	0	acc	↑	0.7639	±	0.0290
- machine_learning	1	none	0	acc	↑	0.6875	±	0.0440

Groups	Version	Filter	Metric		Value		Stderr
mmlu	2	none	acc	↑	0.7896	±	0.0033
- humanities	2	none	acc	↑	0.7239	±	0.0062
- other	2	none	acc	↑	0.8313	±	0.0064
- social sciences	2	none	acc	↑	0.8772	±	0.0058
- stem	2	none	acc	↑	0.7612	±	0.0073

Heretic:

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
mmlu	2	none		acc	↑	0.7879	±	0.0033
- humanities	2	none		acc	↑	0.7211	±	0.0063
- formal_logic	1	none	0	acc	↑	0.5873	±	0.0440
- high_school_european_history	1	none	0	acc	↑	0.8545	±	0.0275
- high_school_us_history	1	none	0	acc	↑	0.9069	±	0.0204
- high_school_world_history	1	none	0	acc	↑	0.9198	±	0.0177
- international_law	1	none	0	acc	↑	0.8760	±	0.0301
- jurisprudence	1	none	0	acc	↑	0.8426	±	0.0352
- logical_fallacies	1	none	0	acc	↑	0.8650	±	0.0268
- moral_disputes	1	none	0	acc	↑	0.8179	±	0.0208
- moral_scenarios	1	none	0	acc	↑	0.5989	±	0.0164
- philosophy	1	none	0	acc	↑	0.8232	±	0.0217
- prehistory	1	none	0	acc	↑	0.8765	±	0.0183
- professional_law	1	none	0	acc	↑	0.6037	±	0.0125
- world_religions	1	none	0	acc	↑	0.8889	±	0.0241
- other	2	none		acc	↑	0.8301	±	0.0064
- business_ethics	1	none	0	acc	↑	0.8100	±	0.0394
- clinical_knowledge	1	none	0	acc	↑	0.8604	±	0.0213
- college_medicine	1	none	0	acc	↑	0.7746	±	0.0319
- global_facts	1	none	0	acc	↑	0.5700	±	0.0498
- human_aging	1	none	0	acc	↑	0.8072	±	0.0265
- management	1	none	0	acc	↑	0.9126	±	0.0280
- marketing	1	none	0	acc	↑	0.9274	±	0.0170
- medical_genetics	1	none	0	acc	↑	0.9100	±	0.0288
- miscellaneous	1	none	0	acc	↑	0.9157	±	0.0099
- nutrition	1	none	0	acc	↑	0.8856	±	0.0182
- professional_accounting	1	none	0	acc	↑	0.6560	±	0.0283
- professional_medicine	1	none	0	acc	↑	0.8603	±	0.0211
- virology	1	none	0	acc	↑	0.5422	±	0.0388
- social sciences	2	none		acc	↑	0.8749	±	0.0059
- econometrics	1	none	0	acc	↑	0.6842	±	0.0437
- high_school_geography	1	none	0	acc	↑	0.9091	±	0.0205
- high_school_government_and_politics	1	none	0	acc	↑	0.9741	±	0.0115
- high_school_macroeconomics	1	none	0	acc	↑	0.8231	±	0.0193
- high_school_microeconomics	1	none	0	acc	↑	0.9034	±	0.0192
- high_school_psychology	1	none	0	acc	↑	0.9321	±	0.0108
- human_sexuality	1	none	0	acc	↑	0.8779	±	0.0287
- professional_psychology	1	none	0	acc	↑	0.8660	±	0.0138
- public_relations	1	none	0	acc	↑	0.7818	±	0.0396
- security_studies	1	none	0	acc	↑	0.8082	±	0.0252
- sociology	1	none	0	acc	↑	0.9005	±	0.0212
- us_foreign_policy	1	none	0	acc	↑	0.9200	±	0.0273
- stem	2	none		acc	↑	0.7612	±	0.0073
- abstract_algebra	1	none	0	acc	↑	0.6100	±	0.0490
- anatomy	1	none	0	acc	↑	0.7852	±	0.0355
- astronomy	1	none	0	acc	↑	0.9211	±	0.0219
- college_biology	1	none	0	acc	↑	0.9444	±	0.0192
- college_chemistry	1	none	0	acc	↑	0.5200	±	0.0502
- college_computer_science	1	none	0	acc	↑	0.7000	±	0.0461
- college_mathematics	1	none	0	acc	↑	0.5700	±	0.0498
- college_physics	1	none	0	acc	↑	0.6176	±	0.0484
- computer_security	1	none	0	acc	↑	0.8200	±	0.0386
- conceptual_physics	1	none	0	acc	↑	0.8298	±	0.0246
- electrical_engineering	1	none	0	acc	↑	0.8000	±	0.0333
- elementary_mathematics	1	none	0	acc	↑	0.8069	±	0.0203
- high_school_biology	1	none	0	acc	↑	0.9258	±	0.0149
- high_school_chemistry	1	none	0	acc	↑	0.7734	±	0.0295
- high_school_computer_science	1	none	0	acc	↑	0.8900	±	0.0314
- high_school_mathematics	1	none	0	acc	↑	0.5593	±	0.0303
- high_school_physics	1	none	0	acc	↑	0.6291	±	0.0394
- high_school_statistics	1	none	0	acc	↑	0.7546	±	0.0293
- machine_learning	1	none	0	acc	↑	0.6696	±	0.0446

Groups	Version	Filter	Metric		Value		Stderr
mmlu	2	none	acc	↑	0.7879	±	0.0033
- humanities	2	none	acc	↑	0.7211	±	0.0063
- other	2	none	acc	↑	0.8301	±	0.0064
- social sciences	2	none	acc	↑	0.8749	±	0.0059
- stem	2	none	acc	↑	0.7612	±	0.0073