Label the benchmark switch by unit rather than by name

Switch between the two benchmarks instead of stacking them

Stop nesting the info button inside the sort button

Score every combination on every task, leaving no blank cells

Publish the remade authorial ensemble and every combination it enters

Land the post-remake refresh pieces held back by the quota

Publish combinations with the combination method chosen on validation

Publish the Sphragis Metre combination grid

Filter each table by how many representations a model combines

Publish KainoBERTa, the architecture control for KainoBERT

Ship the built bundle carrying the KainoBERT rows

Publish the KainoBERT twin of every GreBERTa contender

Colour-code the Representation column by atomic model type

Publish the seven gated Sphragis trio combinations

Publish the complete fifteen-cell Sphragis combination grid

Publish the remaining three-arm pair combinations

Publish the first three double-combination models

Let plot lines select their model

Publish the Mosteller-Wallace character n-gram count models

Bold the best score in each leaderboard column and highlight selected model lines

Tighten leaderboard chart scales

Publish leaderboard score plots

Add task score plots to leaderboards

Compact leaderboard task tables

Publish the siamese SupCon+CE contender

Publish rebuilt atomic model benchmarks

Publish selected character n-gram benchmarks

Score the authorial models on every configuration of both benchmarks

Add the OLMo-1B authorial language models

Reset leaderboard for rebuilt benchmarks

Report the scheme-only floor beneath the sentence results

Deploy fine-tuned GreBerta leaderboard entry

Publish fine-tuned GreBerta results

Publish the syntax-plus-metre arm of the Sphragis-Metre ablation trio

Publish TF-IDF lemma results for Sphragis Metre

Publish metrical line-feature results on the Sphragis-Metre track

Nudge confusion matrix column labels right

Let legend swatches toggle tasks in the per-author chart

Publish TF-IDF lemma logistic regression results

Sort leaderboard tables by 1-sentence score by default

Publish Burrows logistic regression ablations

Correct Burrows Delta classifier label

Merge metre benchmarks and deploy character n-grams

Add Sphragis-Metre character n-grams

Publish Sphragis Metre Burrows Delta results

Deploy Sphragis-Metre syntax results

Add Sphragis-Metre syntax results

Separate dimensions and character n-gram winners

Reframe leaderboard around representations

Expand selected representation text