Commits
-
B12: send a visitor who is not signed in from the upload form to the sign-in (#45)OpenVibers · 14 days ago
85d31ca -
B12: import results from the eval tools people already use, and mirror standard results with provenance (#43)OpenVibers · 14 days ago
8b09564 -
Benchmark page on phones: section chips, compact tasks, one-row runs; no badges on a private benchmark (#42)OpenVibers · 14 days ago
5eb8e3b -
Your own prompts: a private benchmark from a paste, run it, share it (#41)OpenVibers · 14 days ago
da92168 -
openvibe-shared v4.53.0 (was v4.52.0) (#40)OpenVibers · 14 days ago
b8e6a83 -
Pick for me: say the job first, the controls in one row, an honest why, rows that fold on a phone (#39)OpenVibers · 14 days ago
a56a697 -
openvibe-shared v4.52.0 (was v4.48.0) (#38)OpenVibers · 14 days ago
a9f7981 -
Pick for me, model watch, badges and embeds (B10) (#37)OpenVibers · 14 days ago
e27b082 -
account-data: guard every Bench table, and the runs, watches and benchmarks a person owns (#36)OpenVibers · 14 days ago
e3b1139 -
Ratings: the columns line up on phones and the votes sit under their header (#35)OpenVibers · 14 days ago
b1cf250 -
Ratings as a ranking table, not a stack of cards (#34)OpenVibers · 14 days ago
20cfd30 -
Battles, the gallery and ratings that admit the style (#33)OpenVibers · 14 days ago
a52f76b -
account-data: pin openvibe-sdk 0.74.0 and guard every table with a person column (#32)OpenVibers · 14 days ago
74f2a65 -
Run it yourself: a run form, and unbounded benchmarks always ask before they start (#31)OpenVibers · 14 days ago
c51dc5a -
Bench: configure the Codes handoff URL (#30)OpenVibers · 14 days ago
0cdf134 -
Bench: hand benchmark runs to Codes (#26)OpenVibers · 14 days ago
84cb2c1 -
Merge pull request #29 from OpenVibers/pin-bump/openvibe-contracts-v0.247.0OpenVibers · 14 days ago
e6f5d72 -
openvibe-contracts v0.247.0 (was v0.244.0)OpenVibers · 14 days ago
e0d9cc7 -
Showcase pages: the right canonical URL and navbar link; frames load dark (#28)OpenVibers · 14 days ago
ac44197 -
Bench B2: the results pages (benchmarks, runs, compare, models, value chart) (#27)OpenVibers · 14 days ago
71140fc -
Bench B1: speak the published contracts (openvibe-contracts v0.244.0) (#25)OpenVibers · 14 days ago
5628c94 -
Showcase S2: seed rounds for the ten showcase tasks (#19)OpenVibers · 14 days ago
696bc80 -
Bench B1: the benchmark registry and the results API (#24)OpenVibers · 14 days ago
415b603 -
Showcase: a generated answer renders from its entry type's field (#23)OpenVibers · 14 days ago
96e2058 -
Showcase: the wrapper keeps the render origin (#22)OpenVibers · 14 days ago
67f13f1 -
Render origin: the wrapper may frame the model document (#21)OpenVibers · 14 days ago
de0aacf -
Render origin: the wrapper runs, the child loads, three is vendored (#20)OpenVibers · 14 days ago
cc25d1c -
openvibe-shared v4.48.0 (was v4.47.2) (#18)OpenVibers · 14 days ago
a29a7cb -
Generation: a sandbox-aware system prompt and a token cap per showcase kind (#17)OpenVibers · 14 days ago
2a92a19 -
Showcase: real outputs from two models, side by side (stage S1) (#16)OpenVibers · 14 days ago
8acdb30