1.5 KiB
1.5 KiB
Selector
This folder contains rule-based selector configuration and examples.
The v0.2 selector should not call an LLM.
It should use simple matching rules:
- Trigger keywords
- Input types
- Negative triggers
- Pipeline position
- Selection priority
- Negative trigger first
- No-call threshold
- QPI-before-IA gate
- QPI positive-signal gate
- Direct-execution no-call signals
- Calibration smoke checks
Current files:
selector_rules.jsonselector_examples.jsonselector_calibration_inputs.jsonqpi_case_digests.json
The selector does not call an LLM.
The executable demo is scripts/run_selector_demo.py.
Selector regression is scripts/run_selector_regression.py.
Selector calibration smoke check is scripts/run_selector_calibration_smoke.py.
qpi_case_digests.json stores owner-reviewed QPI case digests that can feed selector calibration or later regression selection. It is not a model status upgrade and should not include unreviewed draft cases.
Digest field notes:
misclassification_riskis the canonical digest field matching QPI's structured output contract.qpi_complexity_patternrecords judgment complexity:not_mixed,intra_frame_mixed, orinter_viewpoint_divergence.qpi_complexity_pattern=intra_frame_mixeddoes not implyclassification=mixed;classificationis the final routing class, whileqpi_complexity_patternrecords the structure of the judgment.- Multi-perspective / inter-viewpoint cases require
classification_by_viewpointorviewpoint_summary.