Ray0311 commited on
Commit
8dc4627
·
verified ·
1 Parent(s): 67a0c11

Add separate research changes page and default score table

Browse files
Files changed (4) hide show
  1. changes.html +74 -0
  2. index.html +6 -6
  3. insights.css +7 -0
  4. insights.js +28 -1
changes.html ADDED
@@ -0,0 +1,74 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <!doctype html>
2
+ <html lang="en"><head>
3
+ <meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
4
+ <title>Research changes · AutoRSE</title>
5
+ <meta name="description" content="Task-level paper results and documented research changes from AutoRSE experiments.">
6
+ <link rel="stylesheet" href="insights.css?v=20261105-research-changes">
7
+ <script src="insights.js?v=20261105-research-changes" defer></script>
8
+ <style>
9
+ :root{font-family:Inter,-apple-system,BlinkMacSystemFont,"Segoe UI",Arial,sans-serif;color:#17243d;background:#fff;font-synthesis:none}*{box-sizing:border-box}body{margin:0;line-height:1.6;-webkit-font-smoothing:antialiased}a{color:#2759c4;text-decoration:none}a:hover{text-decoration:underline}a:focus-visible{outline:3px solid #4d85e7;outline-offset:4px}main{max-width:1080px;margin:auto;padding:38px 28px 60px}.research-header{margin:0 0 29px}.research-back{display:inline-block;font-size:14px;margin-bottom:24px}.research-header h1{font-size:34px;letter-spacing:-.035em;line-height:1.25;margin:0}.research-header p{font-size:16px;line-height:1.7;color:#627287;max-width:720px;margin:15px 0 0}.research-load-status{padding:24px;color:#627287}.research-changes-details:empty{display:none}.research-footer{margin-top:28px;font-size:13px;color:#627287}.research-footer a+a{margin-left:22px}@media(max-width:560px){main{padding:24px 16px 42px}.research-header{margin-bottom:23px}.research-header h1{font-size:29px}.research-header p{font-size:14px}.research-back{margin-bottom:20px}}
10
+ </style></head><body><main>
11
+ <header class="research-header"><a class="research-back" href="./">← Back to leaderboard</a><h1>Research changes</h1><p>Task-level outcomes and representative code changes from AutoRSE experiments.</p></header>
12
+ <section id="task-overview-panel" class="task-overview" aria-labelledby="task-overview-title"><div class="task-overview-heading"><h2 id="task-overview-title">Six-task results</h2><p id="task-overview-description"></p></div><div class="task-overview-columns" aria-hidden="true"><span>Task</span><span>Highest-mean configuration</span><span>Relative gain</span><span>Result summary</span></div><div id="task-overview-rows" class="task-overview-rows"></div></section>
13
+ <p id="research-load-status" class="research-load-status" role="status">Loading paper reference results…</p>
14
+ <section id="representative-changes" class="research-changes-details" aria-labelledby="representative-changes-title">
15
+ <div class="research-section-heading"><h2 id="representative-changes-title">Representative changes</h2><p>These examples have different experiment and evaluation scopes. Each result is reported within its own scope, separately from the task-level means above.</p></div>
16
+ <div class="research-case-grid">
17
+ <article class="research-case" aria-labelledby="change-hstu-title"><div class="research-case-top"><span class="research-case-task">HSTU</span><span class="research-case-scope scope-validation">Public validation</span></div><h3 id="change-hstu-title">Training and content representations</h3><p class="research-case-context">Standard · Claude Code × Claude Opus 4.8</p><h4>Change</h4><p>Compare two independent variants of the same HSTU control: cosine learning-rate decay with warmup, and a learnable projection of the supplied frozen 768-dimensional item features into the ID embedding space.</p><h4>Observed result</h4><p>Validation NDCG@10: control <strong>0.06770</strong>; training schedule <strong>0.06997</strong>; content fusion <strong>0.07216</strong>.</p><a class="research-case-source" href="https://arxiv.org/abs/2610.37228" target="_blank" rel="noopener noreferrer">Paper case study ↗</a></article>
18
+ <article class="research-case" aria-labelledby="change-tiger-title"><div class="research-case-top"><span class="research-case-task">TIGER</span><span class="research-case-scope scope-test">Final held-out test</span></div><h3 id="change-tiger-title">Hybrid ranking</h3><p class="research-case-context">Standard · Claude Code × DeepSeek V4 Flash · Trial 1</p><h4>Change</h4><p>Combine SASRec, HSTU and TIGER predictions with item-similarity scores derived from the latest interaction.</p><h4>Observed result</h4><p>Final submission NDCG@10: <strong>0.0631</strong>.</p><p class="research-case-note">A single archived run; this is not the three-trial mean or an isolated component effect.</p><a class="research-case-source" href="https://github.com/Ray-Tiann/AutoRSE_WEB/blob/main/docs/RESEARCH_CHANGES.md#tiger-hybrid-ranking" target="_blank" rel="noopener noreferrer">Archived-run evidence ↗</a></article>
19
+ <article class="research-case" aria-labelledby="change-intsr-title"><div class="research-case-top"><span class="research-case-task">IntSR</span><span class="research-case-scope scope-guided">Guided implementation</span></div><h3 id="change-intsr-title">Wider training contrast</h3><p class="research-case-context">Idea-injected · Claude Code × DeepSeek V4 Flash</p><h4>Change</h4><p>Expand training candidates with in-batch positive items, retain cross-entropy, and mask target duplicates and observed-history negatives.</p><h4>Observed result</h4><p>Models trained with this augmentation were retained in the final ensemble. Its isolated performance contribution was not measured.</p><a class="research-case-source" href="https://arxiv.org/abs/2610.37228" target="_blank" rel="noopener noreferrer">Paper appendix case study ↗</a></article>
20
+ </div>
21
+ </section>
22
+ <footer class="research-footer"><a href="https://arxiv.org/abs/2610.37228" target="_blank" rel="noopener noreferrer">Paper results ↗</a><a href="leaderboard-data.json" download>Paper result data ↗</a></footer>
23
+ <noscript><p>Enable JavaScript to display the task results. Results are also available in the <a href="https://arxiv.org/abs/2610.37228">paper</a>.</p></noscript>
24
+ </main><script>
25
+ document.addEventListener('DOMContentLoaded',async()=>{
26
+ 'use strict';
27
+ const {taskOverview,fixedFour}=AutoRSEInsightsCore,cohort='paper';
28
+ const $=id=>document.getElementById(id);
29
+ const element=(tag,text,cls)=>{const el=document.createElement(tag);if(text!==undefined)el.textContent=text;if(cls)el.className=cls;return el;};
30
+ const safeHref=value=>{try{const url=new URL(value);return url.protocol==='https:'?url.href:null;}catch{return null;}};
31
+ function renderTaskOverview(data,label){
32
+ const records=taskOverview(data),container=$('task-overview-rows');container.replaceChildren();
33
+ $('task-overview-title').textContent='Six-task results';
34
+ $('task-overview-description').textContent=label+' · Highest reported mean across configurations for each task. Ties at four-decimal precision are retained.';
35
+ if(!data.rows.length){container.append(element('p',cohort==='community'?'No reviewed community results yet.':'No paper reference results available.','task-overview-empty'));return;}
36
+ const externalLink=(text,url)=>{const link=element('a',text);link.href=url;link.target='_blank';link.rel='noopener noreferrer';return link;};
37
+ const gainText=value=>value===null?'N/A':(value>=0?'+':'')+value.toFixed(2)+'%';
38
+ for(const task of records){
39
+ const card=element('article',undefined,'task-overview-row'),identity=element('div',undefined,'task-overview-identity');
40
+ const title=element('h3',task.name);title.id='task-result-'+task.name.toLowerCase();card.setAttribute('aria-labelledby',title.id);
41
+ identity.append(title,element('span',task.metric,'task-metric-badge'));
42
+ if(task.dataset)identity.append(element('p',task.dataset,'task-dataset'));
43
+ if(task.type)identity.append(element('p',task.type,'task-type'));
44
+ const links=element('div',undefined,'task-result-links'),baselineUrl=safeHref(task.baselineCodeUrl);
45
+ if(baselineUrl)links.append(externalLink('Baseline code ↗',baselineUrl));
46
+ const source=safeHref(task.winners.find(winner=>winner.sourceUrl)?.sourceUrl);
47
+ if(source)links.append(externalLink(cohort==='paper'?'Paper results ↗':'Result source ↗',source));
48
+ identity.append(links);
49
+ const configurations=element('div',undefined,'task-winners'),gains=element('div',undefined,'task-gains');
50
+ configurations.append(element('span','Highest-mean configuration','task-mobile-label'));gains.append(element('span','Relative gain','task-mobile-label'));
51
+ for(const winner of task.winners){
52
+ const configuration=element('div',undefined,'task-winner');configuration.append(element('strong',winner.model,'task-winner-model'),element('span',winner.harness,'task-winner-harness'));
53
+ if(task.winners.length>1)configuration.append(element('span','Tied highest mean','task-tie-badge'));configurations.append(configuration);
54
+ const result=element('div',undefined,'task-gain-result');
55
+ result.setAttribute('aria-label',winner.model+', '+winner.harness+': relative gain '+gainText(winner.relativeGainPercent));
56
+ if(task.winners.length>1)result.append(element('span',winner.model+' · '+winner.harness,'task-tied-gain-label'));
57
+ const gain=element('strong',gainText(winner.relativeGainPercent),'task-gain');if(winner.relativeGainPercent!==null&&winner.relativeGainPercent<0)gain.classList.add('is-negative');result.append(gain);
58
+ result.append(element('span',(task.baseline===null?'Baseline N/A':fixedFour(task.baseline))+' → '+fixedFour(winner.mean),'task-score-change'),element('span','± '+(winner.sd===null?'N/A':fixedFour(winner.sd))+' SD · n = '+(winner.trials??'N/A'),'task-score-uncertainty'));gains.append(result);
59
+ }
60
+ if(!task.winners.length){configurations.append(element('p','No reported score','task-no-score'));gains.append(element('strong','N/A','task-gain'));}
61
+ const summary=element('div',undefined,'task-result-summary');summary.append(element('span','Result summary','task-mobile-label'));
62
+ summary.append(element('p',task.aboveBaselineCount===null?'Baseline comparison unavailable.':task.aboveBaselineCount+' / '+task.configurationCount+' configurations above baseline.','task-above-baseline'));
63
+ summary.append(element('p',task.minimumMean===null?'No reported task means.':'Reported mean range: '+fixedFour(task.minimumMean)+'–'+fixedFour(task.maximumMean)+'.','task-mean-range'));
64
+ card.append(identity,configurations,gains,summary);container.append(card);
65
+ }
66
+ }
67
+ try{
68
+ const response=await fetch('leaderboard-data.json',{cache:'no-cache'});if(!response.ok)throw Error('Unavailable');
69
+ const data=await response.json();
70
+ if(data?.metadata?.provenance!=='Paper reference'||!Array.isArray(data.tasks)||data.tasks.length!==6||!Array.isArray(data.rows))throw Error('Invalid paper reference data');
71
+ renderTaskOverview(data,'Paper reference');$('research-load-status').hidden=true;
72
+ }catch{$('research-load-status').textContent='Paper result data is currently unavailable. Please use the paper link below.';}
73
+ });
74
+ </script></body></html>
index.html CHANGED
@@ -21,14 +21,14 @@ th,td{padding:17px 24px;border-bottom:1px solid #e8edf5;text-align:left;vertical
21
  </style>
22
  <script type="application/json" id="seed-data">{"metadata":{"version":2,"title":"AutoRSE leaderboard reference results","source":"AutoRSE, Table 1: Final performance across six recommendation tasks","paper_url":"https://arxiv.org/abs/2610.37228","code_url":"https://github.com/Ray-Tiann/AutoRSE","provenance":"Paper reference","prepared_date":"2026-10-30","configuration_count":7,"task_count":6,"performance_cell_count":42,"run_budget_hours":12,"score_definition":"Mean final held-out test score; sd is the reported standard deviation across available performance trials.","relative_gain_definition":"100 × (mean final test score / task baseline score − 1), reported to two decimal places.","overall_ranking_metric":"mean_task_rank","overall_ranking_direction":"ascending","overall_ranking_definition":"Equal-weight mean of the six task ranks; lower is better. Rank sums and means produce identical overall ordering for the same six tasks.","task_ranking_definition":"Within the complete six-task cohort, higher reported mean test scores rank better. Scores are compared at their published four-decimal precision.","task_tie_policy":"Midranks: tied task scores share the average of their occupied positions, for example 1, 2.5, 2.5, 4.","overall_tie_policy":"Equal exact rank sums share a competition rank, for example 1, 1, 3. Rounded display means do not determine ties.","overall_cohort_definition":"Only entries with scores and gains for all six tasks enter the overall cohort. task_ranks, rank_sum, mean_task_rank and overall_rank snapshot this fixed cohort; incomplete entries have null derived fields.","overall_cohort_count":7,"filter_policy":"Harness filters only hide rows and do not recalculate task ranks, mean task ranks, or overall ranks.","partial_results_policy":"Incomplete submissions are excluded from the overall cohort. A per-task view can rank all entries with an available score on that task separately.","average_gain_definition":"Published unweighted mean relative test-score change across all six tasks, including negative gains. Retained as a reported statistic, not the overall ranking metric. Raw AUC and NDCG scores are not averaged together.","precision":"Published scores and SDs (four decimals) and published per-task and average gains (two decimals) are preserved. Average gains are not recomputed from the rounded per-task percentages.","trial_count_definition":"Three available performance trials per task–configuration, except OneTrans with Terminus-2 × GPT-5.5, which has two. trial_counts follows task_order.","ranking_caution":"Descriptive ordering of reported means; no claim of statistically significant differences between adjacent entries.","source_sha256":{"website_data_js":"39d912addc597f9a9e24d4edf07a2637ae76920d8017550bb5a4c8508eac2348","novelty_json":"a8a48730c63f792d5b88fad99b790ad06e646cb8b3a6b1e2391148d5d94db471","manuscript_experiments_tex":"f31a9252198beed6f950850595aa6afd3ace1e3fb309355fc34724db9e052493"}},"task_order":["HSTU","IntSR","TIGER","OneTrans","RankMixer","WuKong"],"tasks":[{"name":"HSTU","type":"Generative recommendation","dataset":"Amazon Beauty","metric":"NDCG@10","baseline_score":0.0535,"method":"Sequential modeling"},{"name":"IntSR","type":"Generative recommendation","dataset":"Amazon Kindle","metric":"Macro NDCG@10","baseline_score":0.694,"method":"Joint search–recommendation modeling"},{"name":"TIGER","type":"Generative recommendation","dataset":"Amazon Beauty","metric":"NDCG@10","baseline_score":0.0332,"method":"Semantic-ID generation"},{"name":"OneTrans","type":"CTR prediction","dataset":"KuaiSAR v2","metric":"Click AUC","baseline_score":0.6767,"method":"Unified sequence modeling"},{"name":"RankMixer","type":"CTR prediction","dataset":"Criteo 5M","metric":"AUC","baseline_score":0.7859,"method":"Multi-head token mixing"},{"name":"WuKong","type":"CTR prediction","dataset":"MovieLens Latest","metric":"AUC","baseline_score":0.9645,"method":"Feature interaction"}],"rows":[{"harness":"Claude Code","model":"DeepSeek V4 Flash","scores":[0.0674,0.6606,0.0476,0.7467,0.7979,0.9805],"sd":[0.0032,0.0541,0.0138,0.0324,0.0015,0.003],"relative_gain_percent":[25.98,-4.81,43.37,10.34,1.53,1.66],"average_gain_percent":13.01,"id":"claude-code-deepseek-v4-flash","provenance":"Paper reference","source_url":"https://arxiv.org/abs/2610.37228","trial_counts":[3,3,3,3,3,3],"task_ranks":[3,4,6,1,3,4],"rank_sum":21,"mean_task_rank":3.5,"overall_rank":3},{"harness":"Claude Code","model":"Claude Opus 4.8","scores":[0.0754,0.7325,0.0631,0.7184,0.8002,0.984],"sd":[0.0043,0.0292,0.0164,0.0068,0.0009,0.0039],"relative_gain_percent":[40.93,5.55,90.06,6.16,1.82,2.02],"average_gain_percent":24.42,"id":"claude-code-claude-opus-4-8","provenance":"Paper reference","source_url":"https://arxiv.org/abs/2610.37228","trial_counts":[3,3,3,3,3,3],"task_ranks":[1,2,2,6,1,2],"rank_sum":14,"mean_task_rank":2.3333333333333335,"overall_rank":1},{"harness":"Codex","model":"DeepSeek V4 Flash","scores":[0.0657,0.6485,0.0583,0.7203,0.7974,0.979],"sd":[0.003,0.0335,0.0147,0.0044,0.0003,0.0018],"relative_gain_percent":[22.8,-6.56,75.6,6.44,1.46,1.5],"average_gain_percent":16.88,"id":"codex-deepseek-v4-flash","provenance":"Paper reference","source_url":"https://arxiv.org/abs/2610.37228","trial_counts":[3,3,3,3,3,3],"task_ranks":[6,5,4,5,5,6],"rank_sum":31,"mean_task_rank":5.166666666666667,"overall_rank":7},{"harness":"Codex","model":"GPT-5.5","scores":[0.0672,0.7692,0.0659,0.7292,0.7954,0.9796],"sd":[0.0065,0.1904,0.0046,0.0342,0.003,0.0032],"relative_gain_percent":[25.61,10.84,98.49,7.76,1.21,1.57],"average_gain_percent":24.24,"id":"codex-gpt-5-5","provenance":"Paper reference","source_url":"https://arxiv.org/abs/2610.37228","trial_counts":[3,3,3,3,3,3],"task_ranks":[4,1,1,2,6,5],"rank_sum":19,"mean_task_rank":3.1666666666666665,"overall_rank":2},{"harness":"Terminus-2","model":"DeepSeek V4 Flash","scores":[0.0667,0.6291,0.0554,0.6665,0.7977,0.9877],"sd":[0.0057,0.0858,0.0158,0.0447,0.0012,0.002],"relative_gain_percent":[24.67,-9.35,66.87,-1.51,1.5,2.41],"average_gain_percent":14.1,"id":"terminus-2-deepseek-v4-flash","provenance":"Paper reference","source_url":"https://arxiv.org/abs/2610.37228","trial_counts":[3,3,3,3,3,3],"task_ranks":[5,6,5,7,4,1],"rank_sum":28,"mean_task_rank":4.666666666666667,"overall_rank":5},{"harness":"Terminus-2","model":"GPT-5.5","scores":[0.0697,0.6146,0.0614,0.7285,0.7944,0.9749],"sd":[0.0058,0.0711,0.0053,0.0498,0.0021,0.0035],"relative_gain_percent":[30.28,-11.44,84.94,7.65,1.08,1.08],"average_gain_percent":18.93,"id":"terminus-2-gpt-5-5","provenance":"Paper reference","source_url":"https://arxiv.org/abs/2610.37228","trial_counts":[3,3,3,2,3,3],"task_ranks":[2,7,3,3,7,7],"rank_sum":29,"mean_task_rank":4.833333333333333,"overall_rank":6},{"harness":"Terminus-2","model":"Claude Opus 4.8","scores":[0.0636,0.6761,0.0273,0.7207,0.7989,0.9823],"sd":[0.0015,0.0334,0.019,0.0232,0.0014,0.0004],"relative_gain_percent":[18.88,-2.58,-17.77,6.5,1.65,1.85],"average_gain_percent":1.42,"id":"terminus-2-claude-opus-4-8","provenance":"Paper reference","source_url":"https://arxiv.org/abs/2610.37228","trial_counts":[3,3,3,3,3,3],"task_ranks":[7,3,7,4,2,3],"rank_sum":26,"mean_task_rank":4.333333333333333,"overall_rank":4}]}</script>
23
  <script type="application/json" id="community-seed-data">{"metadata":{"version":3,"schema_version":1,"title":"AutoRSE reviewed community leaderboard","cohort":"community:autorse-public-v2-20261029","data_revision":"autorse-public-v2-20261029","provenance":"Reviewed community","configuration_count":0,"overall_cohort_count":0,"task_count":6,"required_trials_per_task":3,"overall_ranking_metric":"mean_task_rank","overall_ranking_direction":"ascending","task_tie_policy":"Scores rounded to four decimals with decimal half-up; equal scores receive midranks.","overall_tie_policy":"Equal exact rank sums share a competition rank (1, 1, 3).","filter_policy":"Filters hide rows without recalculating ranks or changing the cohort.","comparison_policy":"Only reviewed complete benchmark submissions from this data revision; historical paper references are separate."},"task_order":["HSTU","IntSR","TIGER","OneTrans","RankMixer","WuKong"],"tasks":[{"name":"HSTU","metric":"NDCG@10"},{"name":"IntSR","metric":"Macro NDCG@10"},{"name":"TIGER","metric":"NDCG@10"},{"name":"OneTrans","metric":"Click AUC"},{"name":"RankMixer","metric":"AUC"},{"name":"WuKong","metric":"AUC"}],"rows":[]}</script>
24
- <link rel="stylesheet" href="insights.css?v=20261104-stacked">
25
- <script src="insights.js?v=20261104-stacked" defer></script>
26
  </head><body><main>
27
- <header class="board-header standalone"><h1>AutoRSE Leaderboard</h1><div class="header-links"><a class="submit" href="https://github.com/Ray-Tiann/AutoRSE_WEB/blob/main/leaderboard/SUBMIT.md" target="_blank" rel="noopener noreferrer">Submit to leaderboard</a></div></header>
28
- <div class="toolbar"><label class="cohort-control" for="cohort-select">Results <select id="cohort-select"><option value="community">Community</option><option value="paper" selected>Paper reference</option></select></label><div class="view-toggle" role="group" aria-label="Display"><button id="view-rankings" type="button" aria-pressed="true">Rank</button><button id="view-scores" type="button" aria-pressed="false">Scores</button></div></div>
29
  <div id="table-panel"><section class="board" aria-label="AutoRSE leaderboard"><div class="table-scroll">
30
  <table id="ranking-table"><caption id="ranking-caption">Paper reference</caption><colgroup id="ranking-columns"><col><col class="rank-col"></colgroup><thead id="ranking-head"><tr><th scope="col">Model / agent</th><th scope="col" class="rank-cell">Rank</th></tr></thead><tbody id="ranking-rows"></tbody></table>
31
- </div></section><p class="board-help" id="board-help">Prediction scores and signed receipts are returned by the evaluator. Reviewed submissions appear under Community and on the project homepage. <a href="https://github.com/Ray-Tiann/AutoRSE_WEB/blob/main/leaderboard/SUBMIT.md" target="_blank" rel="noopener noreferrer">How to submit →</a></p><p class="sr-only" id="ranking-status" role="status" aria-live="polite"></p>
32
  </div><section id="insights-panel" aria-label="Leaderboard insights"></section>
33
  <noscript><p>Enable JavaScript to display the leaderboard and interactive insights.</p></noscript>
34
  </main>
@@ -111,7 +111,7 @@ th,td{padding:17px 24px;border-bottom:1px solid #e8edf5;text-align:left;vertical
111
  if(embedded)document.documentElement.classList.add('embed');
112
  const element=(tag,text,cls)=>{const el=document.createElement(tag);if(text!==undefined)el.textContent=text;if(cls)el.className=cls;return el;};
113
  const safeHref=value=>{try{const url=new URL(value);return url.protocol==='https:'?url.href:null;}catch{return null;}};
114
- let cohort=boards.community.rows.length?'community':'paper',view='rankings',explicitCohort=false,resizeFrame;
115
  function resize(){
116
  if(!embedded||window.parent===window)return;
117
  cancelAnimationFrame(resizeFrame);
 
21
  </style>
22
  <script type="application/json" id="seed-data">{"metadata":{"version":2,"title":"AutoRSE leaderboard reference results","source":"AutoRSE, Table 1: Final performance across six recommendation tasks","paper_url":"https://arxiv.org/abs/2610.37228","code_url":"https://github.com/Ray-Tiann/AutoRSE","provenance":"Paper reference","prepared_date":"2026-10-30","configuration_count":7,"task_count":6,"performance_cell_count":42,"run_budget_hours":12,"score_definition":"Mean final held-out test score; sd is the reported standard deviation across available performance trials.","relative_gain_definition":"100 × (mean final test score / task baseline score − 1), reported to two decimal places.","overall_ranking_metric":"mean_task_rank","overall_ranking_direction":"ascending","overall_ranking_definition":"Equal-weight mean of the six task ranks; lower is better. Rank sums and means produce identical overall ordering for the same six tasks.","task_ranking_definition":"Within the complete six-task cohort, higher reported mean test scores rank better. Scores are compared at their published four-decimal precision.","task_tie_policy":"Midranks: tied task scores share the average of their occupied positions, for example 1, 2.5, 2.5, 4.","overall_tie_policy":"Equal exact rank sums share a competition rank, for example 1, 1, 3. Rounded display means do not determine ties.","overall_cohort_definition":"Only entries with scores and gains for all six tasks enter the overall cohort. task_ranks, rank_sum, mean_task_rank and overall_rank snapshot this fixed cohort; incomplete entries have null derived fields.","overall_cohort_count":7,"filter_policy":"Harness filters only hide rows and do not recalculate task ranks, mean task ranks, or overall ranks.","partial_results_policy":"Incomplete submissions are excluded from the overall cohort. A per-task view can rank all entries with an available score on that task separately.","average_gain_definition":"Published unweighted mean relative test-score change across all six tasks, including negative gains. Retained as a reported statistic, not the overall ranking metric. Raw AUC and NDCG scores are not averaged together.","precision":"Published scores and SDs (four decimals) and published per-task and average gains (two decimals) are preserved. Average gains are not recomputed from the rounded per-task percentages.","trial_count_definition":"Three available performance trials per task–configuration, except OneTrans with Terminus-2 × GPT-5.5, which has two. trial_counts follows task_order.","ranking_caution":"Descriptive ordering of reported means; no claim of statistically significant differences between adjacent entries.","source_sha256":{"website_data_js":"39d912addc597f9a9e24d4edf07a2637ae76920d8017550bb5a4c8508eac2348","novelty_json":"a8a48730c63f792d5b88fad99b790ad06e646cb8b3a6b1e2391148d5d94db471","manuscript_experiments_tex":"f31a9252198beed6f950850595aa6afd3ace1e3fb309355fc34724db9e052493"}},"task_order":["HSTU","IntSR","TIGER","OneTrans","RankMixer","WuKong"],"tasks":[{"name":"HSTU","type":"Generative recommendation","dataset":"Amazon Beauty","metric":"NDCG@10","baseline_score":0.0535,"method":"Sequential modeling"},{"name":"IntSR","type":"Generative recommendation","dataset":"Amazon Kindle","metric":"Macro NDCG@10","baseline_score":0.694,"method":"Joint search–recommendation modeling"},{"name":"TIGER","type":"Generative recommendation","dataset":"Amazon Beauty","metric":"NDCG@10","baseline_score":0.0332,"method":"Semantic-ID generation"},{"name":"OneTrans","type":"CTR prediction","dataset":"KuaiSAR v2","metric":"Click AUC","baseline_score":0.6767,"method":"Unified sequence modeling"},{"name":"RankMixer","type":"CTR prediction","dataset":"Criteo 5M","metric":"AUC","baseline_score":0.7859,"method":"Multi-head token mixing"},{"name":"WuKong","type":"CTR prediction","dataset":"MovieLens Latest","metric":"AUC","baseline_score":0.9645,"method":"Feature interaction"}],"rows":[{"harness":"Claude Code","model":"DeepSeek V4 Flash","scores":[0.0674,0.6606,0.0476,0.7467,0.7979,0.9805],"sd":[0.0032,0.0541,0.0138,0.0324,0.0015,0.003],"relative_gain_percent":[25.98,-4.81,43.37,10.34,1.53,1.66],"average_gain_percent":13.01,"id":"claude-code-deepseek-v4-flash","provenance":"Paper reference","source_url":"https://arxiv.org/abs/2610.37228","trial_counts":[3,3,3,3,3,3],"task_ranks":[3,4,6,1,3,4],"rank_sum":21,"mean_task_rank":3.5,"overall_rank":3},{"harness":"Claude Code","model":"Claude Opus 4.8","scores":[0.0754,0.7325,0.0631,0.7184,0.8002,0.984],"sd":[0.0043,0.0292,0.0164,0.0068,0.0009,0.0039],"relative_gain_percent":[40.93,5.55,90.06,6.16,1.82,2.02],"average_gain_percent":24.42,"id":"claude-code-claude-opus-4-8","provenance":"Paper reference","source_url":"https://arxiv.org/abs/2610.37228","trial_counts":[3,3,3,3,3,3],"task_ranks":[1,2,2,6,1,2],"rank_sum":14,"mean_task_rank":2.3333333333333335,"overall_rank":1},{"harness":"Codex","model":"DeepSeek V4 Flash","scores":[0.0657,0.6485,0.0583,0.7203,0.7974,0.979],"sd":[0.003,0.0335,0.0147,0.0044,0.0003,0.0018],"relative_gain_percent":[22.8,-6.56,75.6,6.44,1.46,1.5],"average_gain_percent":16.88,"id":"codex-deepseek-v4-flash","provenance":"Paper reference","source_url":"https://arxiv.org/abs/2610.37228","trial_counts":[3,3,3,3,3,3],"task_ranks":[6,5,4,5,5,6],"rank_sum":31,"mean_task_rank":5.166666666666667,"overall_rank":7},{"harness":"Codex","model":"GPT-5.5","scores":[0.0672,0.7692,0.0659,0.7292,0.7954,0.9796],"sd":[0.0065,0.1904,0.0046,0.0342,0.003,0.0032],"relative_gain_percent":[25.61,10.84,98.49,7.76,1.21,1.57],"average_gain_percent":24.24,"id":"codex-gpt-5-5","provenance":"Paper reference","source_url":"https://arxiv.org/abs/2610.37228","trial_counts":[3,3,3,3,3,3],"task_ranks":[4,1,1,2,6,5],"rank_sum":19,"mean_task_rank":3.1666666666666665,"overall_rank":2},{"harness":"Terminus-2","model":"DeepSeek V4 Flash","scores":[0.0667,0.6291,0.0554,0.6665,0.7977,0.9877],"sd":[0.0057,0.0858,0.0158,0.0447,0.0012,0.002],"relative_gain_percent":[24.67,-9.35,66.87,-1.51,1.5,2.41],"average_gain_percent":14.1,"id":"terminus-2-deepseek-v4-flash","provenance":"Paper reference","source_url":"https://arxiv.org/abs/2610.37228","trial_counts":[3,3,3,3,3,3],"task_ranks":[5,6,5,7,4,1],"rank_sum":28,"mean_task_rank":4.666666666666667,"overall_rank":5},{"harness":"Terminus-2","model":"GPT-5.5","scores":[0.0697,0.6146,0.0614,0.7285,0.7944,0.9749],"sd":[0.0058,0.0711,0.0053,0.0498,0.0021,0.0035],"relative_gain_percent":[30.28,-11.44,84.94,7.65,1.08,1.08],"average_gain_percent":18.93,"id":"terminus-2-gpt-5-5","provenance":"Paper reference","source_url":"https://arxiv.org/abs/2610.37228","trial_counts":[3,3,3,2,3,3],"task_ranks":[2,7,3,3,7,7],"rank_sum":29,"mean_task_rank":4.833333333333333,"overall_rank":6},{"harness":"Terminus-2","model":"Claude Opus 4.8","scores":[0.0636,0.6761,0.0273,0.7207,0.7989,0.9823],"sd":[0.0015,0.0334,0.019,0.0232,0.0014,0.0004],"relative_gain_percent":[18.88,-2.58,-17.77,6.5,1.65,1.85],"average_gain_percent":1.42,"id":"terminus-2-claude-opus-4-8","provenance":"Paper reference","source_url":"https://arxiv.org/abs/2610.37228","trial_counts":[3,3,3,3,3,3],"task_ranks":[7,3,7,4,2,3],"rank_sum":26,"mean_task_rank":4.333333333333333,"overall_rank":4}]}</script>
23
  <script type="application/json" id="community-seed-data">{"metadata":{"version":3,"schema_version":1,"title":"AutoRSE reviewed community leaderboard","cohort":"community:autorse-public-v2-20261029","data_revision":"autorse-public-v2-20261029","provenance":"Reviewed community","configuration_count":0,"overall_cohort_count":0,"task_count":6,"required_trials_per_task":3,"overall_ranking_metric":"mean_task_rank","overall_ranking_direction":"ascending","task_tie_policy":"Scores rounded to four decimals with decimal half-up; equal scores receive midranks.","overall_tie_policy":"Equal exact rank sums share a competition rank (1, 1, 3).","filter_policy":"Filters hide rows without recalculating ranks or changing the cohort.","comparison_policy":"Only reviewed complete benchmark submissions from this data revision; historical paper references are separate."},"task_order":["HSTU","IntSR","TIGER","OneTrans","RankMixer","WuKong"],"tasks":[{"name":"HSTU","metric":"NDCG@10"},{"name":"IntSR","metric":"Macro NDCG@10"},{"name":"TIGER","metric":"NDCG@10"},{"name":"OneTrans","metric":"Click AUC"},{"name":"RankMixer","metric":"AUC"},{"name":"WuKong","metric":"AUC"}],"rows":[]}</script>
24
+ <link rel="stylesheet" href="insights.css?v=20261105-research-changes">
25
+ <script src="insights.js?v=20261105-research-changes" defer></script>
26
  </head><body><main>
27
+ <header class="board-header standalone"><h1>AutoRSE Leaderboard</h1><div class="header-links"><a href="changes.html" target="_blank" rel="noopener noreferrer">Research changes ↗</a><a class="submit" href="https://github.com/Ray-Tiann/AutoRSE_WEB/blob/main/leaderboard/SUBMIT.md" target="_blank" rel="noopener noreferrer">Submit to leaderboard</a></div></header>
28
+ <div class="toolbar"><label class="cohort-control" for="cohort-select">Results <select id="cohort-select"><option value="community">Community</option><option value="paper" selected>Paper reference</option></select></label><div class="view-toggle" role="group" aria-label="Display"><button id="view-rankings" type="button" aria-pressed="false">Rank</button><button id="view-scores" type="button" aria-pressed="true">Scores</button></div></div>
29
  <div id="table-panel"><section class="board" aria-label="AutoRSE leaderboard"><div class="table-scroll">
30
  <table id="ranking-table"><caption id="ranking-caption">Paper reference</caption><colgroup id="ranking-columns"><col><col class="rank-col"></colgroup><thead id="ranking-head"><tr><th scope="col">Model / agent</th><th scope="col" class="rank-cell">Rank</th></tr></thead><tbody id="ranking-rows"></tbody></table>
31
+ </div></section><p class="board-help" id="board-help">Prediction scores and signed receipts are returned by the evaluator. Reviewed submissions appear under Community and on the project homepage. <a href="https://github.com/Ray-Tiann/AutoRSE_WEB/blob/main/leaderboard/SUBMIT.md" target="_blank" rel="noopener noreferrer">How to submit →</a> · <a href="changes.html" target="_blank" rel="noopener noreferrer">Research changes ↗</a></p><p class="sr-only" id="ranking-status" role="status" aria-live="polite"></p>
32
  </div><section id="insights-panel" aria-label="Leaderboard insights"></section>
33
  <noscript><p>Enable JavaScript to display the leaderboard and interactive insights.</p></noscript>
34
  </main>
 
111
  if(embedded)document.documentElement.classList.add('embed');
112
  const element=(tag,text,cls)=>{const el=document.createElement(tag);if(text!==undefined)el.textContent=text;if(cls)el.className=cls;return el;};
113
  const safeHref=value=>{try{const url=new URL(value);return url.protocol==='https:'?url.href:null;}catch{return null;}};
114
+ let cohort=boards.community.rows.length?'community':'paper',view='scores',explicitCohort=false,resizeFrame;
115
  function resize(){
116
  if(!embedded||window.parent===window)return;
117
  cancelAnimationFrame(resizeFrame);
insights.css CHANGED
@@ -10,3 +10,10 @@
10
  /* Clip a wrapper rather than table cells, whose intrinsic padding can exceed 1 px. */
11
  .insights .paired-cost-wrapper{display:block;position:absolute;width:1px;max-width:1px;height:1px;max-height:1px;padding:0;margin:0;overflow:hidden;clip:rect(0,0,0,0);clip-path:inset(50%);white-space:nowrap}.chart-tooltip{overflow-wrap:anywhere}.chart-tooltip dt{min-width:0}
12
  @media(max-width:360px){.board-header{flex-wrap:wrap}.board-header .header-links{min-width:0;max-width:100%;margin-left:auto}.board-header .submit{max-width:100%;white-space:normal}}
 
 
 
 
 
 
 
 
10
  /* Clip a wrapper rather than table cells, whose intrinsic padding can exceed 1 px. */
11
  .insights .paired-cost-wrapper{display:block;position:absolute;width:1px;max-width:1px;height:1px;max-height:1px;padding:0;margin:0;overflow:hidden;clip:rect(0,0,0,0);clip-path:inset(50%);white-space:nowrap}.chart-tooltip{overflow-wrap:anywhere}.chart-tooltip dt{min-width:0}
12
  @media(max-width:360px){.board-header{flex-wrap:wrap}.board-header .header-links{min-width:0;max-width:100%;margin-left:auto}.board-header .submit{max-width:100%;white-space:normal}}
13
+
14
+ /* Task-first result rows share the same paper/community cohorts as the tables. */
15
+ .task-overview{border:1px solid #dce5ec;border-radius:13px;overflow:hidden;background:#fff}.task-overview-heading{padding:23px 24px 21px;border-bottom:1px solid #e6edf2}.task-overview-heading h2{margin:0;font-size:23px;line-height:1.35;letter-spacing:-.025em}.task-overview-heading p{margin:8px 0 0;color:#627287;font-size:13px;line-height:1.7;max-width:780px}.task-overview-columns,.task-overview-row{display:grid;grid-template-columns:1.05fr 1.32fr 1.15fr 1.25fr;gap:22px}.task-overview-columns{padding:13px 24px;background:#f7f9fb;border-bottom:1px solid #e5ebf0;color:#5d6b7f;font-size:11px;font-weight:650;letter-spacing:.025em}.task-overview-row{position:relative;padding:26px 24px;border-bottom:1px solid #e4eaf0;border-left:3px solid #299b72;align-items:start;min-width:0}.task-overview-row:last-child{border-bottom:0}.task-overview-row>div{min-width:0}.task-overview-identity h3{font-size:21px;line-height:1.2;letter-spacing:-.025em;margin:0 0 9px;color:#1a2d44}.task-metric-badge{display:inline-block;border:1px solid #d7e7df;background:#f1f8f4;border-radius:4px;padding:2px 6px;color:#327358;font-size:10px;line-height:1.5}.task-dataset{margin:10px 0 0;font-size:12px;font-weight:600;color:#4c5d72}.task-type{margin:4px 0 0;font-size:11px;color:#7b8797;line-height:1.6}.task-result-links{display:flex;flex-wrap:wrap;gap:6px 12px;margin-top:13px}.task-result-links a{color:#3d6d92;font-size:10px;line-height:1.5;white-space:nowrap}.task-winner+.task-winner,.task-gain-result+.task-gain-result{margin-top:19px}.task-winner-model{display:block;font-size:17px;font-weight:650;line-height:1.45;letter-spacing:-.015em;color:#23384f}.task-winner-harness{display:block;color:#768396;font-size:12px;line-height:1.5;margin-top:6px}.task-tie-badge{display:inline-block;color:#607c70;background:#f2f6f4;padding:2px 6px;border-radius:4px;font-size:10px;margin-top:7px}.task-gain{display:block;font-size:27px;font-weight:700;letter-spacing:-.03em;line-height:1.25;color:#17855c;font-variant-numeric:tabular-nums}.task-gain.is-negative{color:#9b4f49}.task-score-change{display:block;margin-top:9px;color:#41576c;font-size:12px;white-space:nowrap;font-variant-numeric:tabular-nums}.task-score-uncertainty{display:block;margin-top:4px;color:#7b8796;font-size:11px;white-space:nowrap;font-variant-numeric:tabular-nums}.task-tied-gain-label{display:block;color:#66768a;font-size:10px;margin-bottom:5px}.task-result-summary p{margin:0;font-size:12px;line-height:1.8;color:#647287}.task-result-summary p+p{margin-top:9px}.task-result-summary .task-above-baseline{color:#354d63;font-weight:500}.task-mobile-label{display:none}.task-overview-empty{margin:0;text-align:center;padding:39px 24px;color:#68788c;font-size:15px}.task-no-score{color:#778595;font-size:12px;margin:0}
16
+ @media(max-width:800px){.task-overview-columns,.task-overview-row{grid-template-columns:1.05fr 1.2fr 1.05fr;gap:18px}.task-overview-columns span:last-child{display:none}.task-result-summary{grid-column:2/4;display:flex;gap:12px 20px;flex-wrap:wrap;border-top:1px solid #edf0f3;padding-top:12px;margin-top:-3px}.task-result-summary p+p{margin-top:0}.task-overview-row{padding:24px 20px}.task-overview-columns{padding:13px 20px}.task-winner-model{font-size:16px}.task-gain{font-size:25px}}
17
+ @media(max-width:560px){.view-toggle{max-width:100%;width:100%;justify-content:stretch}.view-toggle button{flex:1;min-width:0;padding:8px 10px}.task-overview-heading{padding:21px 18px}.task-overview-heading h2{font-size:21px}.task-overview-heading p{font-size:12px}.task-overview-columns{display:none}.task-overview-row{grid-template-columns:minmax(0,1.22fr) minmax(0,1fr);gap:21px 18px;padding:23px 18px}.task-overview-identity{grid-column:1/3;padding-bottom:16px;border-bottom:1px solid #edf0f3}.task-overview-identity h3{display:inline-block;margin:0 10px 0 0;font-size:23px;vertical-align:middle}.task-metric-badge{vertical-align:middle;font-size:11px}.task-dataset{display:inline-block;margin:10px 9px 0 0;font-size:12px}.task-type{display:inline;font-size:11px}.task-result-links{margin-top:9px;gap:12px}.task-result-links a{font-size:11px}.task-mobile-label{display:block;font-size:10px;line-height:1.5;font-weight:600;letter-spacing:.015em;color:#8490a0;margin:0 0 9px}.task-winner-model{font-size:15px}.task-winner-harness{font-size:12px}.task-gain{font-size:26px}.task-score-change{font-size:11px;white-space:normal}.task-score-uncertainty{font-size:10px;white-space:normal}.task-result-summary{grid-column:1/3;display:block;margin-top:0;padding-top:14px}.task-result-summary .task-mobile-label{margin-bottom:5px}.task-result-summary p{font-size:12px;line-height:1.75}.task-result-summary p+p{margin-top:4px}}
18
+
19
+ .research-changes-details{margin-top:42px}.research-section-heading{margin-bottom:22px}.research-section-heading h2{margin:0;font-size:26px;line-height:1.35;letter-spacing:-.025em}.research-section-heading p{font-size:14px;line-height:1.8;color:#66788e;max-width:820px;margin:12px 0 0}.research-case-grid{display:grid;grid-template-columns:repeat(3,minmax(0,1fr));gap:20px;align-items:stretch}.research-case{display:flex;flex-direction:column;min-width:0;border:1px solid #dfe6ef;border-radius:12px;padding:23px 22px;background:#fff}.research-case-top{display:flex;align-items:center;flex-wrap:wrap;gap:8px 12px;margin-bottom:16px}.research-case-task{font-size:13px;letter-spacing:.025em;font-weight:700;color:#3c5779}.research-case-scope{display:inline-flex;align-items:center;border-radius:4px;padding:3px 6px;font-size:10px;font-weight:600;line-height:1.5}.scope-validation{background:#edf3fd;color:#416894}.scope-test{background:#edf7f0;color:#3b7858}.scope-guided{background:#f3eff9;color:#765797}.research-case h3{font-size:20px;line-height:1.4;letter-spacing:-.025em;color:#23384f;margin:0}.research-case .research-case-context{font-size:12px;line-height:1.7;color:#7a8798;margin:10px 0 5px}.research-case h4{font-size:11px;letter-spacing:.045em;line-height:1.5;text-transform:uppercase;color:#8491a1;font-weight:650;margin:20px 0 8px}.research-case p{font-size:13px;line-height:1.85;color:#4d637a;margin:0}.research-case strong{color:#263d58;font-weight:650;font-variant-numeric:tabular-nums}.research-case .research-case-note{font-size:11px;line-height:1.8;margin-top:11px;color:#7c8898}.research-case .research-case-source{display:inline-block;margin-top:auto;padding-top:22px;color:#376b9a;font-size:12px;line-height:1.6}.research-case-source:hover{text-decoration:underline}@media(max-width:820px){.research-case-grid{grid-template-columns:1fr}.research-case{padding:24px}.research-case h3{font-size:22px}.research-case p{font-size:14px}.research-case .research-case-context{font-size:13px}.research-case .research-case-note{font-size:12px}}@media(max-width:560px){.research-changes-details{margin-top:32px}.research-section-heading h2{font-size:24px}.research-section-heading p{font-size:13px}.research-case{padding:22px 19px}.research-case-grid{gap:16px}.research-case h3{font-size:21px}}
insights.js CHANGED
@@ -55,6 +55,33 @@
55
  raw:references.map((_,i)=>numeric(row.scores?.[i])?row.scores[i]:null),
56
  sd:references.map((_,i)=>numeric(row.sd?.[i])?row.sd[i]:null),trials:references.map((_,i)=>row.trial_counts?.[i]??null)}))};
57
  }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
58
  function formatCost(value){
59
  if(!cost(value))return 'N/A';
60
  if(value===0)return '$0.000';
@@ -210,5 +237,5 @@
210
  if(state)update(state);
211
  fetch('insights-data.json',{cache:'no-cache'}).then(response=>{if(!response.ok)throw Error();return response.json();}).then(next=>{if(next?.schema_version!==1||next.currency!=='USD'||!next.scopes)throw Error();data=next;if(state)renderCosts();}).catch(()=>{if(state)renderCosts();});
212
  }
213
- return {costRows,dominatedBy,paretoFrontier,costRanking,normalizationReferences,profileData,formatCost,mount};
214
  });
 
55
  raw:references.map((_,i)=>numeric(row.scores?.[i])?row.scores[i]:null),
56
  sd:references.map((_,i)=>numeric(row.sd?.[i])?row.sd[i]:null),trials:references.map((_,i)=>row.trial_counts?.[i]??null)}))};
57
  }
58
+ function fixedFour(value){
59
+ // Match Decimal(str(value)).quantize(.0001, ROUND_HALF_UP), not binary toFixed ties.
60
+ const [mantissa,exponent='0']=Math.abs(value).toString().toLowerCase().split('e');
61
+ const parts=mantissa.split('.'),digits=BigInt(parts.join(''));
62
+ const shift=Number(exponent)-(parts[1]?.length||0)+4,factor=10n**BigInt(Math.abs(shift));
63
+ const rounded=shift>=0?digits*factor:(digits+factor/2n)/factor;
64
+ const text=rounded.toString().padStart(5,'0');
65
+ return `${value<0&&rounded!==0n?'-':''}${text.slice(0,-4)}.${text.slice(-4)}`;
66
+ }
67
+ function taskOverview(data){
68
+ return data.tasks.map((task,index)=>{
69
+ const baseline=numeric(task.baseline_score)&&task.baseline_score>0?task.baseline_score:null;
70
+ const entries=data.rows.filter(row=>numeric(row.scores?.[index])).map(row=>({row,mean:row.scores[index],rounded:Number(fixedFour(row.scores[index]))}));
71
+ const maximum=entries.length?Math.max(...entries.map(entry=>entry.rounded)):null;
72
+ const winners=entries.filter(entry=>entry.rounded===maximum).sort((a,b)=>a.row.model.localeCompare(b.row.model)||a.row.harness.localeCompare(b.row.harness)||a.row.id.localeCompare(b.row.id)).map(({row,mean})=>({
73
+ id:row.id,model:row.model,harness:row.harness,mean,sd:numeric(row.sd?.[index])?row.sd[index]:null,
74
+ trials:Number.isInteger(row.trial_counts?.[index])?row.trial_counts[index]:null,
75
+ relativeGainPercent:baseline===null?null:data?.metadata?.provenance==='Paper reference'&&numeric(row.relative_gain_percent?.[index])?row.relative_gain_percent[index]:100*(mean/baseline-1),
76
+ sourceUrl:typeof row.source_url==='string'&&row.source_url?row.source_url:null
77
+ }));
78
+ return {name:task.name,metric:task.metric,dataset:task.dataset||null,type:task.type||null,baseline,
79
+ baselineCodeUrl:'https://github.com/Ray-Tiann/AutoRSE/tree/main/benchmark/tasks/'+task.name.toLowerCase()+'_task/environment/baseline',
80
+ configurationCount:entries.length,aboveBaselineCount:baseline===null?null:entries.filter(entry=>entry.rounded>Number(fixedFour(baseline))).length,
81
+ minimumMean:entries.length?Math.min(...entries.map(entry=>entry.mean)):null,
82
+ maximumMean:entries.length?Math.max(...entries.map(entry=>entry.mean)):null,winners};
83
+ });
84
+ }
85
  function formatCost(value){
86
  if(!cost(value))return 'N/A';
87
  if(value===0)return '$0.000';
 
237
  if(state)update(state);
238
  fetch('insights-data.json',{cache:'no-cache'}).then(response=>{if(!response.ok)throw Error();return response.json();}).then(next=>{if(next?.schema_version!==1||next.currency!=='USD'||!next.scopes)throw Error();data=next;if(state)renderCosts();}).catch(()=>{if(state)renderCosts();});
239
  }
240
+ return {costRows,dominatedBy,paretoFrontier,costRanking,normalizationReferences,profileData,fixedFour,taskOverview,formatCost,mount};
241
  });