Publish PRISM-Memory adapter bundle
Browse files
README.md
CHANGED
|
@@ -28,6 +28,15 @@ checkpoint from the original `better_memory` project.
|
|
| 28 |
This comparison holds the QA layer constant. It compares extractor against
|
| 29 |
extractor, not a full end-to-end GPT-4.1 system.
|
| 30 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 31 |
## Load the adapter
|
| 32 |
|
| 33 |
```python
|
|
|
|
| 28 |
This comparison holds the QA layer constant. It compares extractor against
|
| 29 |
extractor, not a full end-to-end GPT-4.1 system.
|
| 30 |
|
| 31 |
+
## Why this is useful
|
| 32 |
+
|
| 33 |
+
- It keeps hard limits and preferences available for later workflow generation.
|
| 34 |
+
- It keeps current state separate from future plans.
|
| 35 |
+
- It supports dated recall and clean refusal on unsupported questions.
|
| 36 |
+
|
| 37 |
+
See [docs/release/memory-scenarios.md](docs/release/memory-scenarios.md) for
|
| 38 |
+
the compact end-to-end examples.
|
| 39 |
+
|
| 40 |
## Load the adapter
|
| 41 |
|
| 42 |
```python
|