Show pure Kinex RLE evaluation progress; retain native VLA and historical Kinex evidence

hard/app.js

@@ -116,7 +116,7 @@ function recordingComparisonNote(ep){
 
 function campaignProgress(){
  const c=data.active_campaign;if(!c)return '';const n=c.counts||{};const paired=(n.paired_replacement_pending||0)+(n.paired_replacement_running||0);const codexOnly=(n.codex_replacement_pending||0)+(n.codex_replacement_running||0);const kinex=Object.values(c.kinex_reruns||{});const kinexDone=kinex.filter(x=>x.closed&&x.published_episodes===1).length;
- return `<section class="panel" style="padding:18px 24px;margin-bottom:24px" aria-label="Current evaluation campaign"><strong>${esc(c.phase||'Kinex + RoboEnv · GPT-6 Astra / high')}</strong>${c.dispatch_enabled===false?'<p><strong>New dispatch paused.</strong> '+esc(c.pause_reason||'Runtime review pending.')+'</p>':''}<p>${number(n.finished||0)} / ${number(c.planned_episodes)} episodes completed · ${number(n.running||0)} running · ${number(n.queued||0)} queued${n.preflight?' · '+number(n.preflight)+' checking initialization':''}${paired?' · '+number(paired)+' paired replacements pending':''}${codexOnly?' · '+number(codexOnly)+' Codex replacements pending':''}${n.preflight_failed?' · '+number(n.preflight_failed)+' initialization failures awaiting replacement':''}${n.not_run?' · '+number(n.not_run)+' unavailable':''}${n.needs_review?' · '+number(n.needs_review)+' under review':''}. ${c.closed&&c.published_episodes===n.finished?'Evaluations have finished. All available results, videos and evidence are listed below.':'Results, videos and evidence are published after verification.'}</p>${kinex.length?'<p>'+number(kinexDone)+' / '+number(kinex.length)+' authorized Kinex reruns verified and published. Original selected results remain visible until replacement verification finishes.</p>':''}<p class="source-note">Updated ${esc(c.updated_at)} · <a href="${esc(c.status_url)}" target="_blank" rel="noopener">Campaign status and provenance ↗</a></p></section>`;
+ return `<section class="panel" style="padding:18px 24px;margin-bottom:24px" aria-label="Current evaluation campaign"><strong>${esc(c.phase||'Kinex + RoboEnv · GPT-6 Astra / high')}</strong>${c.dispatch_enabled===false?'<p><strong>New dispatch paused.</strong> '+esc(c.pause_reason||'Runtime review pending.')+'</p>':''}<p>${number(n.finished||0)} / ${number(c.planned_episodes)} episodes completed · ${number(n.running||0)} running · ${number(n.queued||0)} queued${n.preflight?' · '+number(n.preflight)+' checking initialization':''}${paired?' · '+number(paired)+' paired replacements pending':''}${codexOnly?' · '+number(codexOnly)+' Codex replacements pending':''}${n.preflight_failed?' · '+number(n.preflight_failed)+' initialization failures awaiting replacement':''}${n.not_run?' · '+number(n.not_run)+' awaiting a seed decision':''}${n.needs_review?' · '+number(n.needs_review)+' under review':''}. ${c.closed&&c.published_episodes===n.finished?'Evaluations have finished. All available results, videos and evidence are listed below.':'Results, videos and evidence are published after verification.'}</p>${kinex.length?'<p>'+number(kinexDone)+' / '+number(kinex.length)+' authorized Kinex reruns verified and published. Original selected results remain visible until replacement verification finishes.</p>':''}<p class="source-note">Updated ${esc(c.updated_at)} · <a href="${esc(c.status_url)}" target="_blank" rel="noopener">Campaign status and provenance ↗</a></p></section>`;
 }
 function opusPilotProgress(){
  const c=data.pilot_campaigns?.['cc-opus'];if(!c)return '';const n=c.counts||{};
@@ -224,7 +224,7 @@ function overviewDetail(task,ep) {
   const initializationBlocked=(data.active_campaign?.initialization_blocked||[]).includes(task.key+'/'+current.variant);
   const review=reviewFor(task.key,current.variant);
   if(!ep&&review)return `<section class="panel pending-panel" data-runtime-review><span class="badge invalid">Under review</span><h2>Native success recorded; runtime finalization failed</h2><p>${esc(review.reason)}</p><p>Original harness verdict: invalid. Recording manifest: incomplete. This attempt is excluded from formal success metrics; its usage is retained separately.</p><dl class="facts"><dt>Native terminal steps</dt><dd>${number(actionSteps(review.native_terminal_steps,task,current.variant))}</dd><dt>Fresh input</dt><dd>${number(review.usage.fresh_input_tokens)}</dd><dt>Cache read</dt><dd>${number(review.usage.cached_input_tokens)}</dd><dt>Output</dt><dd>${number(review.usage.output_tokens)}</dd><dt>Model responses</dt><dd>${number(review.usage.response_count)}</dd><dt>Estimated API-equivalent cost</dt><dd>${exactMoney(review.estimated_api_equivalent_usd)}</dd></dl><p>Both original camera streams decode successfully. Their completion manifest was not sealed, so they are retained as review evidence.</p><div class="inline-links">${Object.entries(review.links).map(([key,url])=>external(url,key.replaceAll('_',' '))).join('')}</div></section>${instructionPanel(task,null)}`;
-  if(!ep)return `<div class="pending-layout"><section class="pending-panel"><span class="pending-orbit">○</span><span class="eyebrow">${initializationBlocked?'Initialization blocked':'Awaiting evaluation'}</span><h2>${seedLimit?.exhausted?'No native-valid seed at or below '+number(seedLimit.max_seed):initializationBlocked?'Native seed 0 could not initialize':'No recorded episode yet'}</h2><p>${variantName(current.variant)} / ${esc(variantFor(current.variant).effort)}</p><div class="pending-facts">${metric(number(actionSteps(plannedProtocol(task,current.variant).max_control_steps,task)),'Campaign planned action steps')}${metric(String(task.planned_seed_by_variant?.[current.variant]??plannedProtocol(task,current.variant).seed),'Seed')}${metric(plannedProtocol(task,current.variant).control_frequency_hz+' Hz','Control frequency')}</div><p class="subtle">${seedLimit?.exhausted?'Native checks exhausted seeds 1–'+number(seedLimit.max_seed)+'. No Codex model attempt was launched. The original Kinex result is retained; this unavailable Codex task is excluded from success rates.':initializationBlocked?'No model was called. Seed 0 is retained by user decision; this is excluded from task success rates.':'This task is included in the planned catalog. Results and usage will appear after evaluation.'}</p></section>${instructionPanel(task,null)}</div>${seedLimit?.evidence_url?external(seedLimit.evidence_url,'Native seed selection evidence'):''}${nativeNote(task,null)}`;
+  if(!ep)return `<div class="pending-layout"><section class="pending-panel"><span class="pending-orbit">○</span><span class="eyebrow">${initializationBlocked?'Initialization blocked':'Awaiting evaluation'}</span><h2>${seedLimit?.exhausted?'No native-valid seed at or below '+number(seedLimit.max_seed):initializationBlocked?'Native seed 0 could not initialize':'No recorded episode yet'}</h2><p>${variantName(current.variant)} / ${esc(variantFor(current.variant).effort)}</p><div class="pending-facts">${metric(number(actionSteps(plannedProtocol(task,current.variant).max_control_steps,task)),'Campaign planned action steps')}${metric(String(task.planned_seed_by_variant?.[current.variant]??plannedProtocol(task,current.variant).seed),'Seed')}${metric(plannedProtocol(task,current.variant).control_frequency_hz+' Hz','Control frequency')}</div><p class="subtle">${seedLimit?.exhausted?'Native checks exhausted seeds 1–'+number(seedLimit.max_seed)+'. No Codex model attempt was launched. The original Kinex result is retained; this unavailable Codex task is excluded from success rates.':initializationBlocked?'No model was called. The authored seed could not initialize; a seed-selection decision is pending. This case is excluded from native success rates.':'This task is included in the planned catalog. Results and usage will appear after evaluation.'}</p></section>${instructionPanel(task,null)}</div>${seedLimit?.evidence_url?external(seedLimit.evidence_url,'Native seed selection evidence'):''}${nativeNote(task,null)}`;
   const m=ep.media||{},u=ep.usage||{},a=ep.analysis||{};
   const reason=a.evidence_note||a.failure_reason||a.cause||a.observed||a.visible_completion;
   return `<div class="episode-layout"><div>${videoPanel(ep)}<p class="media-note">${m.playback_kind==='official_action_index'?'Head + left wrist + right wrist · 4× official recording, not simulation time. Wrist views update only at inference boundaries. Native 320 × 240 views are upscaled for display; action-limit recordings omit the post-final-action frame.':ep.recording_mode==='native'?'Native camera recording · cached agent-resolution frames · '+(m.speed||1)+'× simulation speed.':ep.provider==='robobridge'&&m.playback_kind==='live_spectator'?'Independent spectator cameras · 720 px high per view · overview + wrist views · '+(m.speed||1)+'× simulation speed. Agent images retain native resolution and orientation.':m.playback_kind==='native_observation_snapshots'?'Native observation snapshots · 10 FPS playback; this is not simulation time.':`${m.source_fps?m.source_fps+' FPS source capture · ':''}${m.speed||1}× simulation speed.`} Session timestamps use wall time and are not synchronized to video frames.</p></div><aside class="result-panel panel"><div class="result-head"><span class="eyebrow">Native verdict</span>${badge(ep)}</div><div class="result-number">${number(actionSteps(ep.steps,ep))}<span> / ${number(actionSteps(ep.measured_protocol?.max_steps,ep))} action steps</span></div><div class="completion-track"><span style="width:${Math.min(100,(actionSteps(ep.steps,ep)||0)/(actionSteps(ep.measured_protocol?.max_steps,ep)||1)*100)}%"></span></div><dl class="facts"><dt>Termination</dt><dd>${esc(ep.verdict?.termination||ep.verdict?.reason||'—')}</dd><dt>Execution</dt><dd>${esc(ep.execution?.status||'—')}</dd><dt>Evidence</dt><dd>${ep.valid?'Valid':'Invalid'}</dd><dt>Simulation time</dt><dd title="${esc(ep.timing?.simulation_time_note||'')}">${ep.timing&&ep.simulation_time_s==null?'Not recorded':elapsed(ep.simulation_time_s)}</dd><dt>Wall time</dt><dd title="${esc(ep.timing?.wall_time_scope||'')}">${elapsed(ep.wall_time_s)}</dd><dt>Estimated cost</dt><dd>${costValue(stats([ep]))}</dd><dt>Seed</dt><dd>${ep.seed}</dd><dt>Agent version</dt><dd>${ep.kinex_version?'Kinex '+esc(ep.kinex_version):ep.codex_version?'Codex '+esc(ep.codex_version):ep.claude_code_version?'Claude Code '+esc(ep.claude_code_version):esc(variantFor(ep.variant).agent)}</dd></dl><div class="result-links">${external(ep.links.verdict,'Native verdict')}${external(ep.links.provenance,'Source identity')}</div></aside></div>${nativeNote(task,ep)}${ep.post_terminal_review?`<aside class="native-note" data-post-terminal-note><span class="note-symbol">!</span><div><strong>Native episode completed; later provider error</strong><p>${esc(ep.post_terminal_review.reason)}</p>${external(ep.links.completion_review,"Terminal completion review")}</div></aside>`:""}${initializationNote?`<aside class="native-note" data-initialization-note><span class="note-symbol">!</span><div><strong>Measured initial-state variation</strong><p>${esc(initializationNote)}</p><a href="${esc(resultURL('initialization-comparison.json'))}" target="_blank">Paired initialization evidence ↗</a></div></aside>`:''}${ep.usage_review?`<aside class="native-note" data-usage-review-note><span class="note-symbol">i</span><div><strong>${esc(ep.usage_review.title||'Usage accounting review')}</strong><p>${esc(ep.usage_review.reason)}</p>${external(ep.links.usage_review,"Usage reconciliation evidence")}</div></aside>`:""}${ep.native_clock_review?`<aside class="native-note" data-native-clock-note><span class="note-symbol">i</span><div><strong>Native physics clock differs from control-frame time</strong><p>${esc(ep.native_clock_review.reason)}</p>${external(ep.links.native_clock_review,"Native clock review")}</div></aside>`:""}${ep.collector_review&&ep.collector_review.schema!=="rlebench/collector-recovery/1"?`<aside class="native-note" data-collector-review-note><span class="note-symbol">i</span><div><strong>Host collector recovered; original model continued</strong><p>${esc(ep.collector_review.reason)}</p><p>Wall time uses the original session timestamps. Lost host setup timing is left unknown.</p><p>Host waiting protection was extended prospectively from 8 to 16 hours for this original execution; the native control budget is unchanged.</p>${external(ep.links.collector_review,"Collector recovery evidence")}</div></aside>`:""}${ep.provenance?.protocol_exception?.same_seed_comparison===false?`<aside class="native-note" data-seed-scope-note><span class="note-symbol">i</span><div><strong>Different evaluation seeds</strong><p>Codex uses seed ${number(ep.seed)} after native expert and initialization checks with no model calls during selection. ${ep.provenance.protocol_exception.max_seed!=null?'Seed selection was capped at '+number(ep.provenance.protocol_exception.max_seed)+'. ':''} Kinex retains its original seed ${number(ep.provenance.protocol_exception.preserved_kinex_seed)} result. Codex ran once; this is not a same-seed comparison.</p>${external(ep.links.seed_selection,'Native seed selection evidence')}</div></aside>`:''}${recordingComparisonNote(ep)}${episodeMetrics(ep)}<div class="detail-two-column">${instructionPanel(task,ep)}<section class="panel episode-analysis"><div class="section-heading"><div><h2>Episode notes</h2><p>Recorded observations and task-specific interpretation.</p></div></div><div class="analysis-body">${reason?`<p>${esc(reason)}</p>`:`<p>${esc(a.outcome||'The native result and original evidence are retained for this episode.')}</p>`}<details><summary>Read recorded analysis</summary>${Object.entries(a).filter(([k])=>k!=='preflight').map(([k,v])=>`<div class="analysis-entry"><span>${esc(k.replaceAll('_',' '))}</span><p>${esc(typeof v==='string'?v:JSON.stringify(v))}</p></div>`).join('')}</details><div class="inline-links">${external(ep.links.analysis,'Analysis JSON')}${external(ep.links.terminal_frame,'Terminal evidence')}</div></div></section></div><div class="detail-launchers"><a href="${taskRoute(task.key,ep.variant,'session')}"><span class="launcher-icon">≡</span><div><strong>Read the full session</strong><span>${number(ep.session_counts?.visible_events)} events · ${number(ep.session_counts?.observed_images)} observations</span></div>${arrow}</a><a href="${taskRoute(task.key,ep.variant,'artifacts')}"><span class="launcher-icon">⌘</span><div><strong>Inspect agent artifacts</strong><span>${ep.resources?.length||0} saved files · workspace & evidence</span></div>${arrow}</a></div>`;

hard/data-source.json

@@ -1,26 +1,26 @@
 {
   "schema": "programvla/dataset-source/1",
   "dataset": "RLE-Bench/programmable-vla",
-  "revision": "57820ea37b4fdfffc612d8ae4f82937a417365ee",
-  "evidence_root": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/",
+  "revision": "8245a4303b0312cef5b28e75c24fccbe89624c72",
+  "evidence_root": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/",
   "files": {
     "data.json": {
-      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/research/kinex-rle-20261104/panels/hard/data.json"
+      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/research/kinex-rle-20261104/panels/hard/data.json"
     },
     "cost-data.json": {
-      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/research/kinex-rle-20261104/panels/hard/cost-data.json"
+      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/research/kinex-rle-20261104/panels/hard/cost-data.json"
     },
     "measured-instructions.json": {
-      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/research/kinex-rle-20261104/panels/hard/measured-instructions.json"
+      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/research/kinex-rle-20261104/panels/hard/measured-instructions.json"
     },
     "sources.json": {
-      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/research/kinex-rle-20261104/panels/hard/sources.json"
+      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/research/kinex-rle-20261104/panels/hard/sources.json"
     },
     "instruction-audit.json": {
-      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/research/kinex-rle-20261104/panels/hard/instruction-audit.json"
+      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/research/kinex-rle-20261104/panels/hard/instruction-audit.json"
     },
     "attempt-history.json": {
-      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/research/kinex-rle-20261104/panels/hard/attempt-history.json"
+      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/research/kinex-rle-20261104/panels/hard/attempt-history.json"
     }
   }
 }

hard/index.html

@@ -18,9 +18,9 @@
 <a class="skip" href="#main">Skip to content</a>
 <header class="site-header"><div class="header-inner"><a class="brand" href="#overview" aria-label="Programmable VLA home"><img src="assets/kinex-logo.svg" width="34" height="34" alt="Kinex"><span>programmable<span class="brand-divider">/</span><span class="brand-benchmark">vla</span></span></a><nav aria-label="Main navigation"><a href="../plus/index.html">Plus</a><a href="../index.html#overview">Easy</a><a href="#overview" data-nav="overview">Hard</a><a href="#tasks" data-nav="tasks">Explore tasks</a><a href="#instructions" data-nav="instructions">Instructions audit</a><a href="#methodology" data-nav="methodology">Protocol</a></nav><a class="header-link" href="https://huggingface.co/spaces/RLE-Bench/programmable-vla" target="_blank" rel="noopener">Hugging Face <span aria-hidden="true">↗</span></a></div></header>
 <main id="main" tabindex="-1"><div class="loading"><span class="spinner"></span>Loading benchmark results…</div></main>
-<footer class="site-footer"><div><img src="assets/kinex-logo.svg" width="22" height="22" alt=""><strong>Programmable VLA</strong><span class="footer-org">by RLE-Bench</span></div><div><a data-result-file="data.json" href="https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/research/kinex-rle-20261104/panels/hard/data.json" target="_blank">Hard results JSON</a><a data-result-file="sources.json" href="https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/b07d88579523eeeefdf7ad2ca14ee65340af5d4e/hard/sources.json" target="_blank">Data sources</a><a href="https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/b07d88579523eeeefdf7ad2ca14ee65340af5d4e/manifest.json" target="_blank">Manifest</a><span id="updated"></span></div></footer>
+<footer class="site-footer"><div><img src="assets/kinex-logo.svg" width="22" height="22" alt=""><strong>Programmable VLA</strong><span class="footer-org">by RLE-Bench</span></div><div><a data-result-file="data.json" href="https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/research/kinex-rle-20261104/panels/hard/data.json" target="_blank">Hard results JSON</a><a data-result-file="sources.json" href="https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/b07d88579523eeeefdf7ad2ca14ee65340af5d4e/hard/sources.json" target="_blank">Data sources</a><a href="https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/b07d88579523eeeefdf7ad2ca14ee65340af5d4e/manifest.json" target="_blank">Manifest</a><span id="updated"></span></div></footer>
 <dialog id="image-viewer" aria-labelledby="image-caption"><div class="viewer-toolbar"><span id="image-caption"></span><a id="image-original" target="_blank" rel="noopener">Open original ↗</a><button class="icon-button" id="close-viewer" aria-label="Close image">×</button></div><div class="viewer-image"><button id="previous-image" class="viewer-arrow" aria-label="Previous image">←</button><img id="large-image" alt="Recorded agent observation"><button id="next-image" class="viewer-arrow" aria-label="Next image">→</button></div><p>Recorded agent observation · Use ← → to browse · Esc to close</p></dialog>
 <div id="toast" role="status" aria-live="polite"></div>
-<noscript><p>This site needs JavaScript. <a data-result-file="data.json" href="https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/research/kinex-rle-20261104/panels/hard/data.json">Download the complete results index.</a></p></noscript>
+<noscript><p>This site needs JavaScript. <a data-result-file="data.json" href="https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/research/kinex-rle-20261104/panels/hard/data.json">Download the complete results index.</a></p></noscript>
 </body>
 </html>

manifest.json

@@ -99,8 +99,8 @@
       "sha256": "a53108594a23b740fbb08623f6fffdda1d73608b552988a9ffda9f0944ae66a1"
     },
     "hard/app.js": {
-      "bytes": 96115,
-      "sha256": "a1317670c7e29a9d24b9b55e55178162694e7b949d9baeec4924da99a1870e93"
+      "bytes": 96176,
+      "sha256": "65b425cdd73418e4e7e307237e24509d55170a06f633ca5fc64b860ee0bd9019"
     },
     "hard/icon.svg": {
       "bytes": 725,
@@ -112,7 +112,7 @@
     },
     "hard/index.html": {
       "bytes": 3630,
-      "sha256": "6929b900ea871c07ba9498696230fc0e7c854b3e9b6d7b4ae5dd11786ce5defc"
+      "sha256": "c0a41478a754e4166b55369ec3262a34242fa30ffd58414a62a04bd3cd362ba1"
     },
     "hard/policy.css": {
       "bytes": 1067,
@@ -124,11 +124,11 @@
     },
     "hard/data-source.json": {
       "bytes": 1587,
-      "sha256": "813be3ca1fd0d8e90d65b7d3ae58f8287a549afab2ab5ab1acd8e57f8ec417bf"
+      "sha256": "8f0f1048a35f69beaf428a0e7ed380bc2cb64780994eebe1f8b1b078a9b5b486"
     },
     "plus/app.js": {
-      "bytes": 98313,
-      "sha256": "f06e15ec26a2ca3b0f39f3963920aecec2bc5adef786e80ac878ae0521f95338"
+      "bytes": 98374,
+      "sha256": "c208f79c4c755ec65fa14d36081de3482e23b69861df53b5aaa1c84890b228f3"
     },
     "plus/icon.svg": {
       "bytes": 725,
@@ -140,7 +140,7 @@
     },
     "plus/index.html": {
       "bytes": 3639,
-      "sha256": "b2b09630078a0517f99452367ed935d83f693d42eeaf9e0154084b9641d61150"
+      "sha256": "aa82945dfef46fdb7b219be195a19960c012297e20f1876ba230270f96038e8a"
     },
     "plus/policy.css": {
       "bytes": 1067,
@@ -152,7 +152,7 @@
     },
     "plus/data-source.json": {
       "bytes": 1587,
-      "sha256": "24fe9346a2dab85f4e1fcfb698c4723de0467ed54c90a11c252d2a843d2a643d"
+      "sha256": "5fbd93aa40d279cf8c106d4c93e2644779485fa8fab0cad47141dc1990d50488"
     },
     "instruction-audit/sources/task03-manifest.yaml": {
       "bytes": 5747,
@@ -255,7 +255,7 @@
   },
   "kinex_rle_dataset": {
     "repo": "RLE-Bench/programmable-vla",
-    "revision": "57820ea37b4fdfffc612d8ae4f82937a417365ee",
+    "revision": "8245a4303b0312cef5b28e75c24fccbe89624c72",
     "prefix": "research/kinex-rle-20261104",
     "scope": "Pure Kinex, RLE revised instructions; Hard 50 selected, Plus 24 selected",
     "complete": false

plus/app.js

@@ -116,7 +116,7 @@ function recordingComparisonNote(ep){
 
 function campaignProgress(){
  const c=data.active_campaign;if(!c)return '';const n=c.counts||{};const paired=(n.paired_replacement_pending||0)+(n.paired_replacement_running||0);const codexOnly=(n.codex_replacement_pending||0)+(n.codex_replacement_running||0);const kinex=Object.values(c.kinex_reruns||{});const kinexDone=kinex.filter(x=>x.closed&&x.published_episodes===1).length;
- return `<section class="panel" style="padding:18px 24px;margin-bottom:24px" aria-label="Current evaluation campaign"><strong>${esc(c.phase||'Kinex + RoboEnv · GPT-6 Astra / high')}</strong>${c.dispatch_enabled===false?'<p><strong>New dispatch paused.</strong> '+esc(c.pause_reason||'Runtime review pending.')+'</p>':''}<p>${number(n.finished||0)} / ${number(c.planned_episodes)} episodes completed · ${number(n.running||0)} running · ${number(n.queued||0)} queued${n.preflight?' · '+number(n.preflight)+' checking initialization':''}${paired?' · '+number(paired)+' paired replacements pending':''}${codexOnly?' · '+number(codexOnly)+' Codex replacements pending':''}${n.preflight_failed?' · '+number(n.preflight_failed)+' initialization failures awaiting replacement':''}${n.not_run?' · '+number(n.not_run)+' unavailable':''}${n.needs_review?' · '+number(n.needs_review)+' under review':''}. ${c.closed&&c.published_episodes===n.finished?'Evaluations have finished. All available results, videos and evidence are listed below.':'Results, videos and evidence are published after verification.'}</p>${kinex.length?'<p>'+number(kinexDone)+' / '+number(kinex.length)+' authorized Kinex reruns verified and published. Original selected results remain visible until replacement verification finishes.</p>':''}<p class="source-note">Updated ${esc(c.updated_at)} · <a href="${esc(c.status_url)}" target="_blank" rel="noopener">Campaign status and provenance ↗</a></p></section>`;
+ return `<section class="panel" style="padding:18px 24px;margin-bottom:24px" aria-label="Current evaluation campaign"><strong>${esc(c.phase||'Kinex + RoboEnv · GPT-6 Astra / high')}</strong>${c.dispatch_enabled===false?'<p><strong>New dispatch paused.</strong> '+esc(c.pause_reason||'Runtime review pending.')+'</p>':''}<p>${number(n.finished||0)} / ${number(c.planned_episodes)} episodes completed · ${number(n.running||0)} running · ${number(n.queued||0)} queued${n.preflight?' · '+number(n.preflight)+' checking initialization':''}${paired?' · '+number(paired)+' paired replacements pending':''}${codexOnly?' · '+number(codexOnly)+' Codex replacements pending':''}${n.preflight_failed?' · '+number(n.preflight_failed)+' initialization failures awaiting replacement':''}${n.not_run?' · '+number(n.not_run)+' awaiting a seed decision':''}${n.needs_review?' · '+number(n.needs_review)+' under review':''}. ${c.closed&&c.published_episodes===n.finished?'Evaluations have finished. All available results, videos and evidence are listed below.':'Results, videos and evidence are published after verification.'}</p>${kinex.length?'<p>'+number(kinexDone)+' / '+number(kinex.length)+' authorized Kinex reruns verified and published. Original selected results remain visible until replacement verification finishes.</p>':''}<p class="source-note">Updated ${esc(c.updated_at)} · <a href="${esc(c.status_url)}" target="_blank" rel="noopener">Campaign status and provenance ↗</a></p></section>`;
 }
 function opusPilotProgress(){
  const c=data.pilot_campaigns?.['cc-opus'];if(!c)return '';const n=c.counts||{};
@@ -224,7 +224,7 @@ function overviewDetail(task,ep) {
   const initializationBlocked=(data.active_campaign?.initialization_blocked||[]).includes(task.key+'/'+current.variant);
   const review=reviewFor(task.key,current.variant);
   if(!ep&&review)return `<section class="panel pending-panel" data-runtime-review><span class="badge invalid">Under review</span><h2>Native success recorded; runtime finalization failed</h2><p>${esc(review.reason)}</p><p>Original harness verdict: invalid. Recording manifest: incomplete. This attempt is excluded from formal success metrics; its usage is retained separately.</p><dl class="facts"><dt>Native terminal steps</dt><dd>${number(actionSteps(review.native_terminal_steps,task,current.variant))}</dd><dt>Fresh input</dt><dd>${number(review.usage.fresh_input_tokens)}</dd><dt>Cache read</dt><dd>${number(review.usage.cached_input_tokens)}</dd><dt>Output</dt><dd>${number(review.usage.output_tokens)}</dd><dt>Model responses</dt><dd>${number(review.usage.response_count)}</dd><dt>Estimated API-equivalent cost</dt><dd>${exactMoney(review.estimated_api_equivalent_usd)}</dd></dl><p>Both original camera streams decode successfully. Their completion manifest was not sealed, so they are retained as review evidence.</p><div class="inline-links">${Object.entries(review.links).map(([key,url])=>external(url,key.replaceAll('_',' '))).join('')}</div></section>${instructionPanel(task,null)}`;
-  if(!ep)return `<div class="pending-layout"><section class="pending-panel"><span class="pending-orbit">○</span><span class="eyebrow">${initializationBlocked?'Initialization blocked':'Awaiting evaluation'}</span><h2>${seedLimit?.exhausted?'No native-valid seed at or below '+number(seedLimit.max_seed):initializationBlocked?'Native seed 0 could not initialize':'No recorded episode yet'}</h2><p>${variantName(current.variant)} / ${esc(variantFor(current.variant).effort)}</p><div class="pending-facts">${metric(number(actionSteps(plannedProtocol(task,current.variant).max_control_steps,task)),'Campaign planned action steps')}${metric(String(task.planned_seed_by_variant?.[current.variant]??plannedProtocol(task,current.variant).seed),'Seed')}${metric(plannedProtocol(task,current.variant).control_frequency_hz+' Hz','Control frequency')}</div><p class="subtle">${seedLimit?.exhausted?'Native checks exhausted seeds 1–'+number(seedLimit.max_seed)+'. No Codex model attempt was launched. The original Kinex result is retained; this unavailable Codex task is excluded from success rates.':initializationBlocked?'No model was called. Seed 0 is retained by user decision; this is excluded from task success rates.':'This task is included in the planned catalog. Results and usage will appear after evaluation.'}</p></section>${instructionPanel(task,null)}</div>${seedLimit?.evidence_url?external(seedLimit.evidence_url,'Native seed selection evidence'):''}${nativeNote(task,null)}`;
+  if(!ep)return `<div class="pending-layout"><section class="pending-panel"><span class="pending-orbit">○</span><span class="eyebrow">${initializationBlocked?'Initialization blocked':'Awaiting evaluation'}</span><h2>${seedLimit?.exhausted?'No native-valid seed at or below '+number(seedLimit.max_seed):initializationBlocked?'Native seed 0 could not initialize':'No recorded episode yet'}</h2><p>${variantName(current.variant)} / ${esc(variantFor(current.variant).effort)}</p><div class="pending-facts">${metric(number(actionSteps(plannedProtocol(task,current.variant).max_control_steps,task)),'Campaign planned action steps')}${metric(String(task.planned_seed_by_variant?.[current.variant]??plannedProtocol(task,current.variant).seed),'Seed')}${metric(plannedProtocol(task,current.variant).control_frequency_hz+' Hz','Control frequency')}</div><p class="subtle">${seedLimit?.exhausted?'Native checks exhausted seeds 1–'+number(seedLimit.max_seed)+'. No Codex model attempt was launched. The original Kinex result is retained; this unavailable Codex task is excluded from success rates.':initializationBlocked?'No model was called. The authored seed could not initialize; a seed-selection decision is pending. This case is excluded from native success rates.':'This task is included in the planned catalog. Results and usage will appear after evaluation.'}</p></section>${instructionPanel(task,null)}</div>${seedLimit?.evidence_url?external(seedLimit.evidence_url,'Native seed selection evidence'):''}${nativeNote(task,null)}`;
   const m=ep.media||{},u=ep.usage||{},a=ep.analysis||{};
   const reason=a.evidence_note||a.failure_reason||a.cause||a.observed||a.visible_completion;
   return `<div class="episode-layout"><div>${videoPanel(ep)}<p class="media-note">${m.playback_kind==='official_action_index'?'Head + left wrist + right wrist · 4× official recording, not simulation time. Wrist views update only at inference boundaries. Native 320 × 240 views are upscaled for display; action-limit recordings omit the post-final-action frame.':ep.recording_mode==='native'?'Native camera recording · cached agent-resolution frames · '+(m.speed||1)+'× simulation speed.':ep.provider==='robobridge'&&m.playback_kind==='live_spectator'?'Independent spectator cameras · 720 px high per view · overview + wrist views · '+(m.speed||1)+'× simulation speed. Agent images retain native resolution and orientation.':m.playback_kind==='native_observation_snapshots'?'Native observation snapshots · 10 FPS playback; this is not simulation time.':`${m.source_fps?m.source_fps+' FPS source capture · ':''}${m.speed||1}× simulation speed.`} Session timestamps use wall time and are not synchronized to video frames.</p></div><aside class="result-panel panel"><div class="result-head"><span class="eyebrow">Native verdict</span>${badge(ep)}</div><div class="result-number">${number(actionSteps(ep.steps,ep))}<span> / ${number(actionSteps(ep.measured_protocol?.max_steps,ep))} action steps</span></div><div class="completion-track"><span style="width:${Math.min(100,(actionSteps(ep.steps,ep)||0)/(actionSteps(ep.measured_protocol?.max_steps,ep)||1)*100)}%"></span></div><dl class="facts"><dt>Termination</dt><dd>${esc(ep.verdict?.termination||ep.verdict?.reason||'—')}</dd><dt>Execution</dt><dd>${esc(ep.execution?.status||'—')}</dd><dt>Evidence</dt><dd>${ep.valid?'Valid':'Invalid'}</dd><dt>Simulation time</dt><dd title="${esc(ep.timing?.simulation_time_note||'')}">${ep.timing&&ep.simulation_time_s==null?'Not recorded':elapsed(ep.simulation_time_s)}</dd><dt>Wall time</dt><dd title="${esc(ep.timing?.wall_time_scope||'')}">${elapsed(ep.wall_time_s)}</dd><dt>Estimated cost</dt><dd>${costValue(stats([ep]))}</dd><dt>Seed</dt><dd>${ep.seed}</dd><dt>Agent version</dt><dd>${ep.kinex_version?'Kinex '+esc(ep.kinex_version):ep.codex_version?'Codex '+esc(ep.codex_version):ep.claude_code_version?'Claude Code '+esc(ep.claude_code_version):esc(variantFor(ep.variant).agent)}</dd></dl><div class="result-links">${external(ep.links.verdict,'Native verdict')}${external(ep.links.provenance,'Source identity')}</div></aside></div>${nativeNote(task,ep)}${ep.post_terminal_review?`<aside class="native-note" data-post-terminal-note><span class="note-symbol">!</span><div><strong>Native episode completed; later provider error</strong><p>${esc(ep.post_terminal_review.reason)}</p>${external(ep.links.completion_review,"Terminal completion review")}</div></aside>`:""}${initializationNote?`<aside class="native-note" data-initialization-note><span class="note-symbol">!</span><div><strong>Measured initial-state variation</strong><p>${esc(initializationNote)}</p><a href="${esc(resultURL('initialization-comparison.json'))}" target="_blank">Paired initialization evidence ↗</a></div></aside>`:''}${ep.usage_review?`<aside class="native-note" data-usage-review-note><span class="note-symbol">i</span><div><strong>${esc(ep.usage_review.title||'Usage accounting review')}</strong><p>${esc(ep.usage_review.reason)}</p>${external(ep.links.usage_review,"Usage reconciliation evidence")}</div></aside>`:""}${ep.native_clock_review?`<aside class="native-note" data-native-clock-note><span class="note-symbol">i</span><div><strong>Native physics clock differs from control-frame time</strong><p>${esc(ep.native_clock_review.reason)}</p>${external(ep.links.native_clock_review,"Native clock review")}</div></aside>`:""}${ep.collector_review&&ep.collector_review.schema!=="rlebench/collector-recovery/1"?`<aside class="native-note" data-collector-review-note><span class="note-symbol">i</span><div><strong>Host collector recovered; original model continued</strong><p>${esc(ep.collector_review.reason)}</p><p>Wall time uses the original session timestamps. Lost host setup timing is left unknown.</p><p>Host waiting protection was extended prospectively from 8 to 16 hours for this original execution; the native control budget is unchanged.</p>${external(ep.links.collector_review,"Collector recovery evidence")}</div></aside>`:""}${ep.provenance?.protocol_exception?.same_seed_comparison===false?`<aside class="native-note" data-seed-scope-note><span class="note-symbol">i</span><div><strong>Different evaluation seeds</strong><p>Codex uses seed ${number(ep.seed)} after native expert and initialization checks with no model calls during selection. ${ep.provenance.protocol_exception.max_seed!=null?'Seed selection was capped at '+number(ep.provenance.protocol_exception.max_seed)+'. ':''} Kinex retains its original seed ${number(ep.provenance.protocol_exception.preserved_kinex_seed)} result. Codex ran once; this is not a same-seed comparison.</p>${external(ep.links.seed_selection,'Native seed selection evidence')}</div></aside>`:''}${recordingComparisonNote(ep)}${episodeMetrics(ep)}<div class="detail-two-column">${instructionPanel(task,ep)}<section class="panel episode-analysis"><div class="section-heading"><div><h2>Episode notes</h2><p>Recorded observations and task-specific interpretation.</p></div></div><div class="analysis-body">${reason?`<p>${esc(reason)}</p>`:`<p>${esc(a.outcome||'The native result and original evidence are retained for this episode.')}</p>`}<details><summary>Read recorded analysis</summary>${Object.entries(a).filter(([k])=>k!=='preflight').map(([k,v])=>`<div class="analysis-entry"><span>${esc(k.replaceAll('_',' '))}</span><p>${esc(typeof v==='string'?v:JSON.stringify(v))}</p></div>`).join('')}</details><div class="inline-links">${external(ep.links.analysis,'Analysis JSON')}${external(ep.links.terminal_frame,'Terminal evidence')}</div></div></section></div><div class="detail-launchers"><a href="${taskRoute(task.key,ep.variant,'session')}"><span class="launcher-icon">≡</span><div><strong>Read the full session</strong><span>${number(ep.session_counts?.visible_events)} events · ${number(ep.session_counts?.observed_images)} observations</span></div>${arrow}</a><a href="${taskRoute(task.key,ep.variant,'artifacts')}"><span class="launcher-icon">⌘</span><div><strong>Inspect agent artifacts</strong><span>${ep.resources?.length||0} saved files · workspace & evidence</span></div>${arrow}</a></div>`;

plus/data-source.json

@@ -1,26 +1,26 @@
 {
   "schema": "programvla/dataset-source/1",
   "dataset": "RLE-Bench/programmable-vla",
-  "revision": "57820ea37b4fdfffc612d8ae4f82937a417365ee",
-  "evidence_root": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/",
+  "revision": "8245a4303b0312cef5b28e75c24fccbe89624c72",
+  "evidence_root": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/",
   "files": {
     "data.json": {
-      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/research/kinex-rle-20261104/panels/plus/data.json"
+      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/research/kinex-rle-20261104/panels/plus/data.json"
     },
     "cost-data.json": {
-      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/research/kinex-rle-20261104/panels/plus/cost-data.json"
+      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/research/kinex-rle-20261104/panels/plus/cost-data.json"
     },
     "measured-instructions.json": {
-      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/research/kinex-rle-20261104/panels/plus/measured-instructions.json"
+      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/research/kinex-rle-20261104/panels/plus/measured-instructions.json"
     },
     "sources.json": {
-      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/research/kinex-rle-20261104/panels/plus/sources.json"
+      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/research/kinex-rle-20261104/panels/plus/sources.json"
     },
     "instruction-audit.json": {
-      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/research/kinex-rle-20261104/panels/plus/instruction-audit.json"
+      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/research/kinex-rle-20261104/panels/plus/instruction-audit.json"
     },
     "attempt-history.json": {
-      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/research/kinex-rle-20261104/panels/plus/attempt-history.json"
+      "url": "https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/research/kinex-rle-20261104/panels/plus/attempt-history.json"
     }
   }
 }

plus/index.html

@@ -18,9 +18,9 @@
 <a class="skip" href="#main">Skip to content</a>
 <header class="site-header"><div class="header-inner"><a class="brand" href="#overview" aria-label="Programmable VLA home"><img src="assets/kinex-logo.svg" width="34" height="34" alt="Kinex"><span>programmable<span class="brand-divider">/</span><span class="brand-benchmark">vla</span></span></a><nav aria-label="Main navigation"><a href="../index.html#overview">Easy</a><a href="../hard/index.html#overview">Hard</a><a href="#overview" data-nav="overview">Plus</a><a href="#tasks" data-nav="tasks">Explore tasks</a><a href="#instructions" data-nav="instructions">Instructions audit</a><a href="#methodology" data-nav="methodology">Protocol</a></nav><a class="header-link" href="https://huggingface.co/spaces/RLE-Bench/programmable-vla" target="_blank" rel="noopener">Hugging Face <span aria-hidden="true">↗</span></a></div></header>
 <main id="main" tabindex="-1"><div class="loading"><span class="spinner"></span>Loading benchmark results…</div></main>
-<footer class="site-footer"><div><img src="assets/kinex-logo.svg" width="22" height="22" alt=""><strong>Programmable VLA</strong><span class="footer-org">by RLE-Bench</span></div><div><a data-result-file="data.json" href="https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/research/kinex-rle-20261104/panels/plus/data.json" target="_blank">Plus results JSON</a><a data-result-file="sources.json" href="https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/b07d88579523eeeefdf7ad2ca14ee65340af5d4e/hard/sources.json" target="_blank">Data sources</a><a href="https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/b2d67a67dd2f9d8e12c1f8db13ba8de309739883/manifest.json" target="_blank">Manifest</a><span id="updated"></span></div></footer>
+<footer class="site-footer"><div><img src="assets/kinex-logo.svg" width="22" height="22" alt=""><strong>Programmable VLA</strong><span class="footer-org">by RLE-Bench</span></div><div><a data-result-file="data.json" href="https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/research/kinex-rle-20261104/panels/plus/data.json" target="_blank">Plus results JSON</a><a data-result-file="sources.json" href="https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/b07d88579523eeeefdf7ad2ca14ee65340af5d4e/hard/sources.json" target="_blank">Data sources</a><a href="https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/b2d67a67dd2f9d8e12c1f8db13ba8de309739883/manifest.json" target="_blank">Manifest</a><span id="updated"></span></div></footer>
 <dialog id="image-viewer" aria-labelledby="image-caption"><div class="viewer-toolbar"><span id="image-caption"></span><a id="image-original" target="_blank" rel="noopener">Open original ↗</a><button class="icon-button" id="close-viewer" aria-label="Close image">×</button></div><div class="viewer-image"><button id="previous-image" class="viewer-arrow" aria-label="Previous image">←</button><img id="large-image" alt="Recorded agent observation"><button id="next-image" class="viewer-arrow" aria-label="Next image">→</button></div><p>Recorded agent observation · Use ← → to browse · Esc to close</p></dialog>
 <div id="toast" role="status" aria-live="polite"></div>
-<noscript><p>This site needs JavaScript. <a data-result-file="data.json" href="https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/57820ea37b4fdfffc612d8ae4f82937a417365ee/research/kinex-rle-20261104/panels/plus/data.json">Download the complete results index.</a></p></noscript>
+<noscript><p>This site needs JavaScript. <a data-result-file="data.json" href="https://huggingface.co/datasets/RLE-Bench/programmable-vla/resolve/8245a4303b0312cef5b28e75c24fccbe89624c72/research/kinex-rle-20261104/panels/plus/data.json">Download the complete results index.</a></p></noscript>
 </body>
 </html>