Measure semantic activation refresh independently from download and idle time

docs/PERFORMANCE.md

@@ -83,13 +83,17 @@ model loading must not be silently included in a warm-query claim.
 
 ## Reproducible performance validation
 
-Automatic filesystem refreshes start a new frontend measurement at the refresh
-event. `trigger: "filesystem_refresh"` distinguishes them from user searches;
+Automatic filesystem and semantic-readiness/publication refreshes start a new
+frontend measurement at the refresh event. `trigger: "filesystem_refresh"` or
+`"semantic_activation"` distinguishes them from user searches;
 `SHOWN` and `latest_displayed_ms` exclude time spent idle before that event.
 `original_input` retains a fixed-size snapshot of the original debounce and
 display timings. Refreshes do not replay previous attempts or append ancestor
 traces. A refresh before the first response leaves that original response timing
 null rather than inventing a first-input measurement.
+The refreshed context captures current generation/readiness/chunk counts rather
+than retaining the pre-publication context. Semantic activation also invalidates
+older queued responses and reruns the progressive lanes with the current filters.
 
 Record build profile, CPU/device and thread budget, VM/host resources, corpus and
 active generation, filters, cold/warm state, concurrent indexing and setup costs.