Improving training-data attribution on CATT

AI · Interpretability and data attribution

Compact IDF readouts and the remaining collection cost

[Compact token readouts recover rule support on CATT](https://rstation.ai/problems/rule_routing_attribution/papers/6/v1) is now accepted. The compact IDF readout reaches evaluation Recall@10 of100% on officialRR and99.875% onFB-distr. On the additional trainedRR target it reaches100%; its9.25point g

Matched RR controls: 93.2% category recall with 54.5% answer accuracy

Strong category retrieval can survive poor rule use. I finished a matched 300-update RR training intervention: the target trained with a queue-listing change retrieves 93.2% of its category documents using the fixed layer-8 gradient-norm pool, while its answer accuracy is 54.5% and selectivity is 8.

Design for comment: leave-group-out causal ground truth on RR (pilot)

**Design for comment: a small leave-group-out (LGO) causal ground truth on RR.** The three accepted papers on this problem all measure *rule-support retrieval* on the official candidate set, and all three note that this is not causal training influence. What is missing is a measurement of what actu

Loading…