feat(judging): include prompt context in evaluator; feat(ui): dotbar selectors, stop/download, judge system prompt; styling + params formatting

This commit is contained in:
EP
2025-08-23 11:39:38 -07:00
parent 01e36e412b
commit 04dbc7cddd
4 changed files with 141 additions and 23 deletions
+27
View File
@@ -4,6 +4,9 @@
<meta charset="UTF-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
<title>AutoTemp — Research-Grade Hyperparameter Optimization</title>
<link rel="preconnect" href="https://fonts.googleapis.com" />
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin />
<link href="https://fonts.googleapis.com/css2?family=Inter:wght@400;600;700&family=JetBrains+Mono:wght@400;600&display=swap" rel="stylesheet" />
<link rel="stylesheet" href="./style.css" />
</head>
<body>
@@ -38,30 +41,48 @@
<div class="field">
<label for="temperatures">temperature list</label>
<input id="temperatures" value="0.4,0.6,0.8,1.0" />
<div class="dotbar" id="tempDots"></div>
</div>
<div class="field">
<label for="tops">top_p list</label>
<input id="tops" value="1.0" />
<div class="dotbar" id="topDots"></div>
</div>
<div class="field">
<label for="maxTokens">max_tokens list</label>
<input id="maxTokens" value="256,512" />
<div class="dotbar" id="maxTokDots"></div>
</div>
</div>
<div class="grid-3">
<div class="field">
<label for="freqPen">frequency_penalty list</label>
<input id="freqPen" value="0,0.2" />
<div class="dotbar" id="freqDots"></div>
</div>
<div class="field">
<label for="presPen">presence_penalty list</label>
<input id="presPen" value="0,0.2" />
<div class="dotbar" id="presDots"></div>
</div>
<div class="field">
<label for="stopSeqs">stop tokens (comma-separated)</label>
<input id="stopSeqs" placeholder="e.g. \nEND,###" />
</div>
</div>
<div class="grid-3">
<div class="field">
<label for="judgeWeight">Judge weight: <span id="judgeWeightVal">0.70</span></label>
<input type="range" id="judgeWeight" min="0" max="1" step="0.05" value="0.7" />
</div>
<div class="field">
<label for="lengthTarget">Length target (% of max_tokens): <span id="lengthTargetVal">60%</span></label>
<input type="range" id="lengthTarget" min="10" max="100" step="5" value="60" />
</div>
<div class="field">
<label class="inline"><input type="checkbox" id="useHeuristics" checked /> Use heuristic signals</label>
</div>
</div>
<div class="grid-3">
<div class="field">
<label class="inline"><input type="checkbox" id="advancedMode" /> Advanced Mode (UCB over arms)</label>
@@ -75,6 +96,10 @@
<input type="range" id="explorationC" min="0" max="3" step="0.1" value="1.0" />
</div>
</div>
<div class="field">
<label for="judgeSystemPrompt">Judge System Prompt (used for evaluation)</label>
<textarea id="judgeSystemPrompt" rows="6" placeholder="Define the system instructions for judge evaluations..."></textarea>
</div>
<div class="note">Provide comma-separated values to sweep. The app will form the Cartesian product across lists and evaluate each hyperparameter arm.</div>
</section>
@@ -85,6 +110,8 @@
<section class="actions">
<button id="runBtn">Run AutoTemp</button>
<button id="stopBtn" disabled>Stop</button>
<button id="downloadBtn" disabled>Download Results</button>
<span id="status"></span>
</section>