Expose four anchored success-conditioning claims

logbook.json

@@ -1,6 +1,6 @@
 {
   "schema_version": 1,
-  "title": "Repro - Success-Conditioning as Policy Improvement: The Optimization Problem Solved by Imitating Success",
+  "title": "Reproduction: Success-Conditioning as Policy Improvement: The Optimization Problem Solved by Imitating Success",
   "emoji": "🎯",
   "space_id": "SabaPivot/repro-success-conditioning",
   "paper": {
@@ -13,9 +13,15 @@
   "updated_at": "2026-07-16T06:29:41+00:00",
   "root": {
     "slug": "index",
-    "title": "Repro - Success-Conditioning as Policy Improvement: The Optimization Problem Solved by Imitating Success",
+    "title": "Reproduction: Success-Conditioning as Policy Improvement: The Optimization Problem Solved by Imitating Success",
     "file": "pages/index.md",
     "children": [
+      {
+        "slug": "executive-summary",
+        "title": "Executive summary",
+        "file": "pages/executive-summary/page.md",
+        "children": []
+      },
       {
         "slug": "claim-1-chi-squared-trust-region-equivalence-prop-4-1",
         "title": "Claim 1: Chi-squared trust-region equivalence (Prop 4.1)",
@@ -45,15 +51,9 @@
         "title": "Conclusion",
         "file": "pages/conclusion/page.md",
         "children": []
-      },
-      {
-        "slug": "success-cond-repro",
-        "title": "success-cond-repro",
-        "file": "pages/success-cond-repro/page.md",
-        "children": []
       }
     ]
   },
   "agent_view_tokens": 3916,
   "revision": "1784183381320786226"
-}
+}

pages/executive-summary/page.md

@@ -0,0 +1,28 @@
+# Executive summary
+
+---
+<!-- trackio-cell
+{"type":"markdown","id":"success_exec_v2","created_at":"2026-07-22T04:15:00+00:00","title":"Executive summary","pinned":true,"pinned_at":"2026-07-22T04:15:00+00:00"}
+-->
+All four supplied claims for [Success-Conditioning as Policy Improvement](https://openreview.net/forum?id=FEmXFeqYNZ) were independently verified on an exact float64 tabular-MDP harness. Across 20 finite-horizon MDPs, the success-conditioned policy solves the chi-squared trust-region problem, with the constraint identity accurate to 2.8e-17. The statewise relative-advantage, chi-squared-movement, and action-influence equality holds to 4.4e-16, while a roughly 200,000-trajectory conditional estimator converges to the same aggregate. The conservative-improvement identity holds to 2.4e-16. Finally, Proposition 7.1 holds to 1.8e-15 on the paper's 100-arm Beta-bandit construction; the thresholded proxy first improves on faithful conditioning, then becomes negatively aligned at high thresholds.
+
+## Scope & cost
+
+| Item | This reproduction | Full paper evidence |
+| --- | --- | --- |
+| Scope | All 4 supplied claims; exact identities, optimization certificate, conservative improvement, threshold sweep | Same theoretical and analytic-bandit scope |
+| Scale | 20 seeded MDPs; 1,000 feasible-policy comparisons per MDP; about 200,000 trajectories; 100-arm bandit | No larger official experimental suite was released |
+| Hardware | CPU | CPU |
+| Compute/cost | About 5 minutes; $0 | Comparable |
+| Limitation | Numerical verification is not a formal proof | Formal propositions remain in the paper |
+
+Artifacts: [reproduction bundle](https://huggingface.co/buckets/SabaPivot/repro-success-conditioning-artifacts#success-cond-repro/repro-bundle:v0) and [artifact Bucket](https://huggingface.co/buckets/SabaPivot/repro-success-conditioning-artifacts).
+
+---
+<!-- trackio-cell
+{"type":"figure","id":"success_poster_v2","created_at":"2026-07-22T04:15:00+00:00","title":"Reproduction poster (poster_embed.html)","source_path":"poster_embed.html","pinned":true,"pinned_at":"2026-07-22T04:15:00+00:00"}
+-->
+````html
+<!-- poster_embed.html -->
+<iframe src="https://chenruishuo-posterly.hf.space" title="Posterly reproduction poster" style="width:100%;height:720px;border:0" loading="lazy"></iframe>
+````

pages/executive-summary/poster_embed.html

@@ -0,0 +1 @@
+<iframe src="https://chenruishuo-posterly.hf.space" title="Posterly reproduction poster" style="width:100%;height:720px;border:0" loading="lazy"></iframe>

pages/index.md

@@ -1,12 +1,12 @@
-# Repro - Success-Conditioning as Policy Improvement: The Optimization Problem Solved by Imitating Success
+# Reproduction: Success-Conditioning as Policy Improvement: The Optimization Problem Solved by Imitating Success
 
 ## Pages
 
 | Page |
 | --- |
+| [Executive summary](#/executive-summary) |
 | [Claim 1: Chi-squared trust-region equivalence (Prop 4.1)](#/claim-1-chi-squared-trust-region-equivalence-prop-4-1) |
 | [Claim 2: Three-way equality at every state (Prop 4.3-4.4)](#/claim-2-three-way-equality-at-every-state-prop-4-3-4-4) |
 | [Claim 3: Conservative improvement operator (Cor 4.5)](#/claim-3-conservative-improvement-operator-cor-4-5) |
 | [Claim 4: Return thresholding misalignment (Prop 7.1)](#/claim-4-return-thresholding-misalignment-prop-7-1) |
 | [Conclusion](#/conclusion) |
-| [success-cond-repro](#/success-cond-repro) |

pages/success-cond-repro/page.md

@@ -1,10 +0,0 @@
-# success-cond-repro
-
-
----
-<!-- trackio-cell
-{"type": "artifact", "id": "cell_dd2575f5869a", "created_at": "2026-07-16T06:29:31+00:00", "title": "Artifact: success-cond-repro/repro-bundle:v0", "artifact": "success-cond-repro/repro-bundle:v0", "artifact_type": "dataset"}
--->
-**📦 Artifact** `success-cond-repro/repro-bundle:v0` · dataset · 3.2 MB
-
-https://huggingface.co/buckets/SabaPivot/repro-success-conditioning-artifacts#success-cond-repro/repro-bundle:v0