95 lines
3.8 KiB
Markdown
95 lines
3.8 KiB
Markdown
# 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.
|