Klexikon stat processing

[dennlinger]

Klexikon has some of the fewest issues, and I already thoroughly investigated the remaining samples:

• [x] Check for samples of duplicates: Can be tracked through the versioning changes in v0.3. Basically had some issues with the disambiguation process introducing reference duplicates. Minor issue with 2-3 samples where the summary was longer.
• [x] Compute stats on the dataset
• [x] Compute novelty on dataset