Classifying with small models · Module C of 6
Making a decision
How do scores become a confident answer, or an honest "not sure"?
From vectors to a label: prototypes, a small learned head, softmax, temperature, blending and per-class thresholds.
Topics in this module
Topics
0/1 done🔒 Unlocks at 4 of 4 topics in Phase 1 — 0 done.- C1
Prototypes Planned
Average a class’s example texts into one vector; descriptions act as weights.
- C2
The classifier head Planned
Logistic regression on frozen vectors, with learned weights per class.
- C3
Softmax and temperature
6 minGuidedSoftmax turns scores into probabilities that sum to one; temperature decides how sure those probabilities sound, and calibration makes them honest.
- C4
Combining two scorers Planned
Blend calibrated probabilities; the blend weight shifts as labelled data grows.
- C5
Thresholds Planned
Per-class confidence bars decide whether to answer at all.
- C6
Frozen encoder vs fine-tuning Planned
Train the smallest part that works.