Use model_key as the addressable identifier and wire comparison-index sidecar

Merge remote-tracking branch 'origin/main' into feat/use-new-backend-data

Tighten eval cards UI and clean up stale local data

Drop input_modalities/output_modalities from MODEL_CARD_COLUMNS

Integrate with test backend data

Merge corpus dashboard into home as paper-aligned landing

Deploy DuckDB-backed frontend to

Add DuckDB shadow-read backend with source-metadata fix

Separate policy and researcher views

Add interpretive signals, corpus dashboard, and slice browser

Preserve evaluator_relationship when flattening model hierarchy

improve ux

Differentiate audience modes and tighten eval navigation

Aggregate setup aliases and clarify benchmark variants

Fix RewardBench2 key normalization for matrix leaderboard routing

Improve eval/model UX, lite data paths, and leaderboard clarity

Add per-benchmark comparison histograms on model detail

Improve eval score displays and summary fallbacks

Refresh eval cards UI and backend data flow

Add survey submission and update survey text for public use

fix bugs

fix bugs

ux changes

Add survey

fix: align reporting cues and developer slugs

feat: refine model and benchmark exploration

redesigned

fix data

Refactor: Update benchmarks with realistic data, fix UI stats, and improve About page

new ux

fixed a lot of bugs, centralized schema

unit tests added

added all the new files