# D 巨人方舟通用模型层

## 1. Scope

```text
layer_id: D
function: upgrade Giant Cognition 2.0 into Giant Ark as a formal general model
primary_nodes: D1, D2, D3, D4, D5, D6
```

`巨人方舟` 是正式模型名。模型建成后，原 `巨人认知 2.0` 下线。

## 2. Model Definition

巨人方舟是一套以认知主权为核心、以现实反馈为校准、以行动实验为生成、以身心底盘和环境审计为约束的通用认知治理架构。

## 3. Confirmed Components

```text
D1 巨人认知 2.0 -> 巨人方舟:
  formal replacement and upgrade

D2 GL0:
  body regulation, emotional tolerance, nervous-system state

D3 GL-R:
  relationship/environment layer; prevents social-system waste heat from being internalized as personal fault

D4 GL3:
  QPI as problem-representation tool

D5 GL4:
  metacognition as coordinator, not dictator

D6 intention-action-feedback loop:
  Reality Lab as action module
```

## 4. Source Material Layer

Primary blocks:

| block_id | role |
| --- | --- |
| R04-B001 | Giant Cognition / upgrade entry |
| R04-B002 | model transition material |
| R04-B003 | CBT and theories as modules |
| R04-B004 | GL0 need through self-blame/vulnerability |
| R04-B005 | ACT placement inside architecture |
| R04-B006 | tolerance window and body-state checks |
| R04-B007 | explicit upgrade requirements |
| R05-B001 | formalization entry |
| R05-B003 | Giant Ark definition |
| R06-B003 | action timing and guardrails |

Cross-links:

```text
C4: methodology scheduling
E3: concrete-domain model calling
```

## 5. Distortion Guard

Do not split Giant Cognition, Giant Ark, and Reality Lab into peer models. The confirmed relation is:

```text
Giant Cognition 2.0 -> Giant Ark formal model
Reality Lab -> action module inside intention-action-feedback
```

## 6. Material Extraction

| material_id | source_blocks | type | extracted material | downstream use |
| --- | --- | --- | --- | --- |
| D-M01 | R04-B007, R05-B003 | model upgrade | Giant Ark upgrades Giant Cognition 2.0 by adding GL0 regulation, GL-R environment, action experiment, feedback calibration, and GL4 limitation. | Giant Ark spec |
| D-M02 | R04-B004, R04-B006 | GL0 material | The model must check body regulation, emotional tolerance, nervous-system state, and tolerance window before high-level cognition takes command. | GL0 design |
| D-M03 | R04-B003, R04-B005 | method library | CBT, ACT, and related theories are modules called by the architecture, not architectures that own the whole problem. | theory-calling policy |
| D-M04 | R05-B003 | formal definition | Giant Ark is a general cognitive governance architecture centered on cognitive sovereignty, reality feedback, action experiment, bodily/emotional base, and environment audit. | canonical definition |
| D-M05 | R05-B002, R05-B005 | application value | The methodology discussion defines why Giant Ark matters: it decides which theory/model/tool enters which layer under which trust conditions. | model-card value proposition |
| D-M06 | R06-B003, R07-B003 | action module | Reality Lab belongs inside intention-action-feedback: it turns explanation into small, reversible, observable experiments. | action protocol |
| D-M07 | R02-B003, R03-B004 | layer discipline | Dominant layer and metacognitive limits must be identified before intervention; GL4 coordinates, it does not dominate. | layer-selection rule |

## 7. Reusable Claims

```text
- Giant Ark replaces Giant Cognition 2.0 as the formal model once completed.
- A general model earns application value only if it can call concrete theories and local models without being captured by them.
- Reality feedback is not an afterthought; it is the mechanism that turns explanation into intervention.
```

## 8. Open Use Notes

This layer is ready for a downstream `Giant Ark formal model specification`. It still needs a dedicated model-card pass before canonical publication.