diff --git a/.github/workflows/deploy.yml b/.github/workflows/deploy.yml index 9957b4f..de8b4d1 100644 --- a/.github/workflows/deploy.yml +++ b/.github/workflows/deploy.yml @@ -26,6 +26,10 @@ jobs: - name: Build run: bun run build working-directory: site + env: + # Optional arena backend (see worker/README.md); empty = local-only votes. + VITE_ARENA_API: ${{ vars.VITE_ARENA_API }} + VITE_TURNSTILE_SITE_KEY: ${{ vars.VITE_TURNSTILE_SITE_KEY }} - uses: actions/configure-pages@v5 - uses: actions/upload-pages-artifact@v3 with: diff --git a/README.en.md b/README.en.md index 0a261d1..6023459 100644 --- a/README.en.md +++ b/README.en.md @@ -37,6 +37,12 @@ Pending: Mistral Medium 3.5 · GLM-5.1 · Qwen3.7-Max (PRs welcome) > This table is a hand-maintained snapshot; the site's registry section is generated live from `models/*.json` and is authoritative. +## Arena blind battles & community board + +- Each case defaults to **Arena mode**: two anonymous artifacts battle (A / B / Neither) and identities reveal only after you vote; switch to **Demo Show** to browse everything (tabs are grouped by model — a model with several harnesses can be viewed singly or side by side). +- Casting 03 votes in each case unlocks the **community leaderboard**, ranked by Bradley-Terry scores (the LMSYS Chatbot Arena approach); "neither" counts separately as both-bad. +- Votes stay in the browser by default; deploying `worker/` (Cloudflare Worker + D1 + Turnstile, free tier) upgrades the board to sitewide crowdsourced data — see [worker/README.md](./worker/README.md). + ## Structure ``` diff --git a/README.md b/README.md index f55dce3..bafc558 100644 --- a/README.md +++ b/README.md @@ -37,6 +37,12 @@ NAGI STUDIO 的 LLM 测评案例集:同一段提示词,不同「模型 × Ha > 此表为人工维护的快照;站点上的「模型阵容」表由 `models/*.json` 实时生成,以站点为准。 +## Arena 盲评与社区榜单 + +- 每个案例默认进入 **Arena 模式**:随机抽两份产出匿名对战(A / B / 都不好),投完票才揭晓身份;右上角可切到 **全部查看**(按模型分组浏览,同一模型多个 Harness 可单看或并排对比)。 +- 在两个案例各投满 03 票后解锁**社区榜单**:按 Bradley-Terry 评分排名(LMSYS Chatbot Arena 同思路),「都不好」单独计入双败。 +- 投票默认存浏览器本地;部署 `worker/`(Cloudflare Worker + D1 + Turnstile 防刷,免费档即可)后自动升级为全站众包数据,步骤见 [worker/README.md](./worker/README.md)。 + ## 结构 / Structure ```