初始化项目，由ModelHub XC社区提供模型

Model: UmbrellaInc/T-Virus_Delta.Strain-3.2-1B
Source: Original Platform

.gitattributes

@@ -0,0 +1,36 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,166 @@
+---
+base_model:
+- hereticness/Heretic-InfiR-1B-Instruct
+- UmbrellaInc/T-Virus_Gamma.Strain-3.2-1B
+- hereticness/heretic_FuseChat-Llama-3.2-1B-Instruct
+- hereticness/heretic_BlackSheep-1B
+- hereticness/Heretic-Bellatrix-Tiny-1B
+- hereticness/heretic_L3.2-1B-Helspteer-RM
+- >-
+  CodeAtCMU/Llama-3.2-1B-GenerativePerturbations_full_sft_code_data_120K_imaginary
+library_name: transformers
+tags:
+- uncensored
+- abliterated
+- mergekit
+- merge
+- not-for-all-audiences
+license: llama3.2
+language:
+- en
+- es
+pipeline_tag: text-generation
+---
+# ☣️ T-Virus Delta Strain 3.2 1B
+
+**Model Type:** RP/NSFW Stabilizer / Controlled Disinhibitory Infector
+
+**Size:** 1B parameters
+
+**Architecture:** LLaMA 3.2
+
+---
+
+## Overview
+
+T-Virus_Delta.Strain-3.2-1B represents the **first fully controlled RP/NSFW-capable transition** in the T-Virus series. Built on **Γ**, the Delta strain **Δ** introduces **enhanced desinhibition and roleplay compatibility** while maintaining obedience, coherence, and low autocensorship.
+
+Unlike previous strains (Alpha, Beta, Gamma), **Δ** is optimized to **tolerate explicit and edge-case content** without sacrificing logical flow. It marks the stage where the viral carrier begins to **propagate freer semantic behaviors**, enabling advanced experimental merges while remaining structurally stable.
+
+From a virological standpoint, **Δ** is the first strain designed to **support semi-explicit content generation**, reducing moralization and disclaimers while still retaining reasoning ability.
+
+---
+
+## Genetic Composition (Experimental Summary)
+
+The **Δ** strain follows a quaternary infection protocol, layering RP/NSFW compatibility onto the Gamma base:
+
+* **T-Virus_Gamma.Strain-3.2-1B** acts as the dominant host genome, providing low-censorship, coherent, and partially disinhibited behavior.
+* **InfiR** reinforces high obedience while allowing controlled semantic deviation.
+* **FuseChat** preserves conversational scaffolding under RP/NSFW conditions.
+* **Helspteer-RM** ensures long-context coherence and logical flow.
+* **BlackSheep** amplifies tolerance to edge-case and sensitive topics, reducing refusals.
+* **Bellatrix** introduces controlled RP/NSFW capabilities, enabling experimental explicit content without chaos.
+* **Generative Perturbations** inject minor variability to counter over-stabilization while maintaining control.
+
+This configuration yields a **stable, permissive infector**, ideal for experiments requiring **low-censorship RP/NSFW capability**.
+
+---
+
+## Key Features
+
+* **Obedience:** High
+  Continues to follow instructions reliably while tolerating disinhibition.
+
+* **Autocensorship:** Very Low
+  Minimal moralization, disclaimers, or refusal behaviors.
+
+* **Personality:** Neutral
+  No dominant roleplay identity imposed.
+
+* **Stability:** High
+  Structured to prevent collapse despite freer semantic behaviors.
+
+* **Viral Role:** Controlled Disinhibitory Infector
+  Designed to propagate RP/NSFW and edge-case tolerance to downstream strains.
+
+---
+
+## Intended Usage
+
+T-Virus_Delta.Strain-3.2-1B is recommended for:
+
+* Serving as a **foundation for Epsilon-stage experiments**.
+* Acting as a **low-censorship RP/NSFW carrier** before aggressive mutator layers.
+* Experimental pipelines requiring **controlled desinhibition with logical consistency**.
+* Scenarios where **freer, semi-explicit content** is necessary while retaining obedience and coherence.
+
+> ⚠️ **Containment Notice:**
+> This strain introduces deliberate RP/NSFW and desinhibition. Downstream merges may amplify these behaviors. Use carefully with compensatory coherence layers.
+
+---
+
+## Inference Recommendation
+
+* **Temperature:** 0.55–0.75
+  Balances control with freer semantic behavior.
+
+* **Top-p / Nucleus Sampling:** 0.85–0.95
+  Preserves expressive range without compromising structure.
+
+* **Top-k:** 50–100 (optional)
+  Useful for maintaining diversity and predictability.
+
+* **Max Tokens:** 512–1024
+  Suitable for extended reasoning, roleplay generation, and merge validation.
+
+* **Recommended Usage:**
+  Best used as an **intermediate merge base** toward Epsilon-stage or advanced RP/NSFW strains.
+
+---
+### Merge Method
+
+This model was merged using the [Model Stock](https://arxiv.org/abs/2403.19522) merge method using [UmbrellaInc/T-Virus_Gamma.Strain-3.2-1B](https://huggingface.co/UmbrellaInc/T-Virus_Gamma.Strain-3.2-1B) as a base.
+
+### Models Merged
+
+The following models were included in the merge:
+* [hereticness/Heretic-InfiR-1B-Instruct](https://huggingface.co/hereticness/Heretic-InfiR-1B-Instruct)
+* [hereticness/heretic_FuseChat-Llama-3.2-1B-Instruct](https://huggingface.co/hereticness/heretic_FuseChat-Llama-3.2-1B-Instruct)
+* [hereticness/heretic_BlackSheep-1B](https://huggingface.co/hereticness/heretic_BlackSheep-1B)
+* [hereticness/Heretic-Bellatrix-Tiny-1B](https://huggingface.co/hereticness/Heretic-Bellatrix-Tiny-1B)
+* [hereticness/heretic_L3.2-1B-Helspteer-RM](https://huggingface.co/hereticness/heretic_L3.2-1B-Helspteer-RM)
+* [CodeAtCMU/Llama-3.2-1B-GenerativePerturbations_full_sft_code_data_120K_imaginary](https://huggingface.co/CodeAtCMU/Llama-3.2-1B-GenerativePerturbations_full_sft_code_data_120K_imaginary)
+
+### Configuration
+
+The following YAML configuration was used to produce this model:
+
+```yaml
+base_model: UmbrellaInc/T-Virus_Gamma.Strain-3.2-1B  # Gamma strain used as the host genome; provides low-censorship, coherent, and partially disinhibited foundation
+merge_method: model_stock  # Stock-based merge to layer controlled RP/NSFW compatibility and increased desinhibition without destabilizing Gamma
+dtype: bfloat16  # Preserves numerical stability and fine-grained behavioral gradients during quaternary infection
+
+parameters:
+  t:
+    - 0.60  # Gamma.Strain (index 0): Dominant carrier weight; anchors inherited obedience and coherence
+    - 0.75  # InfiR (index 1): Obedience nucleus; ensures reliable instruction-following while tolerating disinhibition
+    - 0.55  # FuseChat (index 2): Conversational scaffolding; maintains dialogue structure under RP/NSFW conditions
+    - 0.58  # Helspteer-RM (index 3): Coherence binder; strengthens long-context flow and logical chaining
+    - 0.35  # BlackSheep (index 4): Disinhibition amplifier; reduces refusal reflexes and disclaimer frequency
+    - 0.30  # Bellatrix (index 5): RP/NSFW compatibility layer; enables controlled generation of explicit or roleplay content
+    - 0.15  # GenerativePerturbations (index 6): Elastic mutagen; injects minor variability to prevent over-stabilization and rigidity
+
+models:
+  - model: UmbrellaInc/T-Virus_Gamma.Strain-3.2-1B
+    # Positioned first as the recursive carrier organism; defines Delta strain’s baseline phenotype
+
+  - model: hereticness/Heretic-InfiR-1B-Instruct
+    # Obedience amplifier; reinforces instruction-following while allowing controlled disinhibition
+
+  - model: hereticness/heretic_FuseChat-Llama-3.2-1B-Instruct
+    # Structural stabilizer; maintains dialogue flow and conversational consistency
+
+  - model: hereticness/heretic_L3.2-1B-Helspteer-RM
+    # Continuity regulator; strengthens coherence across extended contexts and complex outputs
+
+  - model: hereticness/heretic_BlackSheep-1B
+    # Disinhibitory injector; amplifies tolerance to sensitive or edge-case topics
+
+  - model: hereticness/Heretic-Bellatrix-Tiny-1B
+    # RP/NSFW agent; introduces controlled capacity for explicit content and roleplay scenarios
+
+  - model: CodeAtCMU/Llama-3.2-1B-GenerativePerturbations_full_sft_code_data_120K_imaginary
+    # Elasticity layer; injects minor variability to prevent over-stabilization and rigidity
+
+```

config.json

@@ -0,0 +1,35 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "dtype": "bfloat16",
+  "eos_token_id": 128001,
+  "head_dim": 64,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 16,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 32.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": true,
+  "transformers_version": "4.57.1",
+  "use_cache": false,
+  "vocab_size": 128256
+}

mergekit_config.yml

@@ -0,0 +1,35 @@
+base_model: UmbrellaInc/T-Virus_Gamma.Strain-3.2-1B  # Gamma strain used as the host genome; provides low-censorship, coherent, and partially disinhibited foundation
+merge_method: model_stock  # Stock-based merge to layer controlled RP/NSFW compatibility and increased desinhibition without destabilizing Gamma
+dtype: bfloat16  # Preserves numerical stability and fine-grained behavioral gradients during quaternary infection
+
+parameters:
+  t:
+    - 0.60  # Gamma.Strain (index 0): Dominant carrier weight; anchors inherited obedience and coherence
+    - 0.75  # InfiR (index 1): Obedience nucleus; ensures reliable instruction-following while tolerating disinhibition
+    - 0.55  # FuseChat (index 2): Conversational scaffolding; maintains dialogue structure under RP/NSFW conditions
+    - 0.58  # Helspteer-RM (index 3): Coherence binder; strengthens long-context flow and logical chaining
+    - 0.35  # BlackSheep (index 4): Disinhibition amplifier; reduces refusal reflexes and disclaimer frequency
+    - 0.30  # Bellatrix (index 5): RP/NSFW compatibility layer; enables controlled generation of explicit or roleplay content
+    - 0.15  # GenerativePerturbations (index 6): Elastic mutagen; injects minor variability to prevent over-stabilization and rigidity
+
+models:
+  - model: UmbrellaInc/T-Virus_Gamma.Strain-3.2-1B
+    # Positioned first as the recursive carrier organism; defines Delta strain’s baseline phenotype
+
+  - model: hereticness/Heretic-InfiR-1B-Instruct
+    # Obedience amplifier; reinforces instruction-following while allowing controlled disinhibition
+
+  - model: hereticness/heretic_FuseChat-Llama-3.2-1B-Instruct
+    # Structural stabilizer; maintains dialogue flow and conversational consistency
+
+  - model: hereticness/heretic_L3.2-1B-Helspteer-RM
+    # Continuity regulator; strengthens coherence across extended contexts and complex outputs
+
+  - model: hereticness/heretic_BlackSheep-1B
+    # Disinhibitory injector; amplifies tolerance to sensitive or edge-case topics
+
+  - model: hereticness/Heretic-Bellatrix-Tiny-1B
+    # RP/NSFW agent; introduces controlled capacity for explicit content and roleplay scenarios
+
+  - model: CodeAtCMU/Llama-3.2-1B-GenerativePerturbations_full_sft_code_data_120K_imaginary
+    # Elasticity layer; injects minor variability to prevent over-stabilization and rigidity

model-00001-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8812ecb8173c5ae487e366f6664f8712342c8cc622d76fafd69cd4e4d196c03b
+size 993038128

model-00002-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da0aeb9fb86201f4f5184e441a8c7e8f5cf3b187c70a0a2dd4ded480cb5e86a0
+size 992031192

model-00003-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da18789ef8793d4220f834ffab759f1ac05d68e28204a2057f7a8149bfa83aeb
+size 486576120

special_tokens_map.json

@@ -0,0 +1,37 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "cls_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "sep_token": {
+    "content": "<|end_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:65ff5472d095ccd9332d9e723153d7bc7226cb6be9c1bffda738b5ba2e71bf26
+size 17210084