Turn selected guidance into a qualified learning target

Choose the recurring operation, not merely a file with a favorable score. Here a and b are alternative suppliers of the same interpretation. Retain one while developing that interpretation; keep the current range read and arithmetic execution. Recover a useful action at a history where the intended learner has the inputs that warrant it. Qualify the action against the unit definition and required report. A guide may help a teacher produce that target; its presence does not qualify every resulting action.

SIRI v1, §3 offers one selective construction: induce guidance from interaction experience, compare assisted and unassisted group returns, and use positively credited guided actions as targets for the policy without that guidance. Its group benefit covers the retrieved bundle. The student still needs the actual decision history and observations. Skill0.5 v1, §3 offers a hybrid: retain particular guidance while using task difficulty to choose guided demonstrations or outcome learning. An empty set of successful demonstrations supplies no imitation target. Choose the actual loss and probability domain through 26.4.5; a truncated distribution or implementation-specific reward needs its own meaning.

The finite fixture uses milliseconds and minutes equally often. For milliseconds, output 1 selects division by 1,000 and output 0 leaves the amount unchanged. For minutes, output 1 leaves the amount unchanged and output 0 selects multiplication by 60. Qualified targets are therefore (1,0). The actual numeric amount remains an input to the retained arithmetic operation. This representation fixes which interpretation can change; new unit conventions need a new applicable definition or another representation.

Let p_j = sigmoid(theta_j) be the probability of output 1 in input class j. With equally weighted qualified targets y=(1,0), minimize

L = -[log(p_0) + log(1-p_1)]/2, with dL/dtheta_j = (p_j-y_j)/2.

Starting at theta=(-ln 4, ln 4), both interpretations have probability 0.20 of the correct action. Four updates with step size eta=4 give:

The exact rates follow from this finite action model and its assumed reliable remaining operations. The fourth update meets the stipulated 0.85 floor without a or b. It remains below the unchanged aided performer’s supplied 0.90 rate. Preserve both arrangements for the whole-result comparison; satisfying a floor does not settle whether a loss of accuracy is worth a saving in support. The trained performer with a guide is another joint arrangement to assess.

The target also determines which trace positions enter the loss. Consider a trace containing a user’s request, the assistant’s current-rule call, the returned rule and the assistant’s report:

daVinci-Dev, ICML 2026, §§2–3 uses full-sequence mid-training followed by action-masked fine-tuning. Such a target switch changes what the fit means. A failed attempt can supply a recorded sequence without supplying actions to imitate as successful conduct. Retain whether its context was observed during the attempt or reconstructed with hindsight. A predicted rule remains a prediction; the current-report task consumes an actual current rule. The finite update here fits the two selected interpretation actions.