Yep - that's my understanding.
AFAIK, when a container in Azure Cosmos DB spans multiple physical partitions, multiple inverted indexes may be created. Each physical partition will maintain its own local index. This way, f a query involves an indexed field (like one using an inverted index), and the query cannot leverage the partition key, Cosmos DB will perform an index lookup on each partition to retrieve the results.
If there are two physical partitions, Cosmos DB will perform two separate inverted index searches—one per partition—then merge the results.
If the above response helps answer your question, remember to "Accept Answer" so that others in the community facing similar issues can easily find the solution. Your contribution is highly appreciated.
hth
Marcin