logbook.json

{
 "schema_version": 1,
 "title": "Repro: Recovering Policy-Induced Errors: Benchmarking and Trajectory Synthesis for Robust GUI Agents (GUI-RobustEval + RoTS)",
 "emoji": "\ud83c\udfaf",
 "space_id": "ParetoOptimal/repro-rots",
 "paper": null,
 "tags": [],
 "updated_at": "2026-07-29T04:30:00+00:00",
 "root": {
  "slug": "index",
  "title": "Repro: Recovering Policy-Induced Errors: Benchmarking and Trajectory Synthesis for Robust GUI Agents (GUI-RobustEval + RoTS)",
  "file": "pages/index.md",
  "children": [
   {
    "slug": "paper-and-approach",
    "title": "Paper and approach",
    "file": "pages/paper-and-approach/page.md",
    "children": []
   },
   {
    "slug": "claim-1-mini-gui-robusteval",
    "title": "Claim 1 - Mini GUI-RobustEval (1,216 executable cases)",
    "file": "pages/claim-1-mini-gui-robusteval/page.md",
    "children": []
   },
   {
    "slug": "claim-2-error-depth-metrics",
    "title": "Claim 2 - Four error depths, post-error success + awareness",
    "file": "pages/claim-2-error-depth-metrics/page.md",
    "children": []
   },
   {
    "slug": "claim-3-rots-co-expansion",
    "title": "Claim 3 - RoTS explore-recovery co-expansion",
    "file": "pages/claim-3-rots-co-expansion/page.md",
    "children": []
   },
   {
    "slug": "claim-4-osworld-sota-audit",
    "title": "Claim 4 - RoTS-32B OSWorld 47.4% SOTA (audit-only)",
    "file": "pages/claim-4-osworld-sota-audit/page.md",
    "children": []
   },
   {
    "slug": "claim-5-fde-eir-ablation",
    "title": "Claim 5 - FDE + EIR beat parallel sampling (Table 4 ablation)",
    "file": "pages/claim-5-fde-eir-ablation/page.md",
    "children": []
   },
   {
    "slug": "claim-6-scaling-analysis",
    "title": "Claim 6 - Scaling expansion rounds and dataset size (Figure 5)",
    "file": "pages/claim-6-scaling-analysis/page.md",
    "children": []
   }
  ]
 },
 "agent_view_tokens": 7757,
 "revision": "1785296871036370700"
}