| --- |
| license: apache-2.0 |
| tags: |
| - browser-ai |
| - capability-handoff |
| - tool-use |
| - structured-output |
| - continual-learning |
| --- |
| |
| # Raw KV-Cache Handoff Lab |
|
|
| This model-companion repository defines `KVC1`, a fail-closed container for raw |
| key/value cache tensor bytes, plus an explicit translation contract for moving a |
| contained idea into another model family's cache geometry. |
|
|
| KV tensors are runtime activations, not model weights. Direct reuse requires |
| matching model/tokenizer identity, tensor geometry, RoPE semantics, layout, and |
| sequence position. Cross-family handoff requires a trained and evaluated |
| translator. |
|
|
| The current acceptance run uses deterministic synthetic payload bytes. It |
| validates the KVC1 container and failure boundaries, not live cache extraction |
| or reinjection. Source-runtime extraction and same-family reload are the next |
| integration gates. |
|
|
| The optional `DOMCAP1` sidecar pack, **Analytics Hotshot**, contains a compact expert briefing, |
| trusted host-tool identifiers, a source policy, and evaluation probes. The |
| portfolio fetches the JSON artifact after a visitor elects to load the pack, |
| validates every requested tool against its local allowlist, and injects the |
| bounded briefing into the next request. DuckDB-Wasm—not the language model—runs |
| the calculation. |
|
|
| ## What this is |
|
|
| - A reproducible raw KV-cache serialization and compatibility prototype. |
| - A contract for a versioned source-to-destination translation layer. |
| - A small, inspectable companion artifact for a browser-hosted base model. |
| - A way to measure task improvement against byte and estimated token cost. |
|
|
| ## What this is not |
|
|
| - It is not a fine-tuned checkpoint or a copy of the base model. |
| - The current reference does not yet train a cross-family translator. |
| - It does not prove that a small model inherits a larger model's intelligence. |
| - It grants no credentials, arbitrary network access, SQL execution, or source-code execution. |
|
|
| ## Artifacts |
|
|
| - `reference/kvcache.py`: KVC1 reference reader/writer and inspector. |
| - `translation/bridge-contract.schema.json`: cross-family translator receipt. |
| - `capability-packs/analytics-hotshot.domcap.json`: optional operating sidecar. |
|
|
| The associated schema is `capability-packs/domcap.schema.json`. A host must |
| still maintain its own trusted tool catalog; a remote container must never be |
| allowed to supply executable implementations. |
|
|
| ## Base model and context accounting |
|
|
| The current desktop base model is |
| [`onnx-community/Qwen2.5-0.5B-Instruct`](https://huggingface.co/onnx-community/Qwen2.5-0.5B-Instruct). |
| Its configuration advertises 32,768 maximum positions. The portfolio uses a |
| separate conservative bound of 3,200 dynamic prompt characters for browser |
| reliability and reports the capability pack's estimated prompt footprint. |
|
|
| ## Research status |
|
|
| Version 0.1.0 is a working infrastructure prototype. The included probes define |
| the intended behavior, but comparative base-versus-container capability scores |
| have not yet been published. Until those evaluations exist, the pack's status |
| remains `prototype`, not `validated`. |
|
|
| Two later experiments are specified in the portfolio research note: |
|
|
| 1. directed conflict retention for targeted catastrophic-interference tests; |
| 2. anisotropic relational retrieval, inspired by view-dependent Gaussian |
| representations but not assumed to share their geometry or compression law. |
|
|
