diff --git a/wizard-work/ledger/CHECK-arxiv-tt20 b/wizard-work/ledger/CHECK-arxiv-tt20
new file mode 100644
index 0000000..de924a8
--- /dev/null
+++ b/wizard-work/ledger/CHECK-arxiv-tt20
@@ -0,0 +1,45 @@
+#!/usr/bin/env bash
+# CHECK-arxiv-tt20 -- arxiv-tt20 extract (Extract lane, 2026-10-29).
+# Re-fetches the arXiv export API and replays the parse; diffs the
+# (arxiv_id, title) pairs against the committed arxiv-tt20.json.
+# Exit code IS the verdict: 0 = PASS, nonzero = FAIL.
+set -u
+cd /root/troupe-standup/wizard-work/ledger || exit 1
+
+WORK="$(mktemp -d)" || exit 1
+trap 'rm -rf "$WORK"' EXIT
+
+python3 - "$WORK" <<'PY'
+import json, re, sys, urllib.request
+
+api = ("https://export.arxiv.org/api/query"
+ "?search_query=all:%22test-time%22+OR+all:%22test-time+scaling%22"
+ "+OR+all:%22test_time+compute%22&start=0&max_results=20")
+req = urllib.request.Request(api, headers={"User-Agent": "Mozilla/5.0"})
+xml = urllib.request.urlopen(req, timeout=30).read().decode("utf-8", errors="replace")
+
+got = []
+for entry in re.finditer(r"(.*?)", xml, re.S):
+ b = entry.group(1)
+ iid = re.search(r"(.*?)", b, re.S)
+ tt = re.search(r"(.*?)", b, re.S)
+ if not iid or not tt:
+ print("FAIL: malformed entry in replay")
+ sys.exit(1)
+ got.append((iid.group(1).split("/abs/", 1)[-1].strip(),
+ tt.group(1).strip()))
+
+committed = json.load(open("arxiv-tt20.json", encoding="utf-8"))
+exp = [(p["arxiv_id"], p["title"]) for p in committed["papers"]]
+
+if len(got) != len(exp):
+ print(f"FAIL: count replay={len(got)} committed={len(exp)}")
+ sys.exit(1)
+for i, ((g1, g2), (e1, e2)) in enumerate(zip(got, exp), 1):
+ if g1 != e1 or g2 != e2:
+ print(f"FAIL: entry {i} replay=({g1!r},{g2!r}) committed=({e1!r},{e2!r})")
+ sys.exit(1)
+
+print(f"PASS: {len(exp)} papers replay id+title match committed arxiv-tt20.json")
+PY
+exit $?
diff --git a/wizard-work/ledger/README-arxiv-tt20.md b/wizard-work/ledger/README-arxiv-tt20.md
new file mode 100644
index 0000000..a338e2e
--- /dev/null
+++ b/wizard-work/ledger/README-arxiv-tt20.md
@@ -0,0 +1,38 @@
+# arxiv-tt20 (Extract, 2026-10-29)
+
+Data extraction and table for the 20 most recent arXiv papers on test-time
+compute / test-time scaling, fetched from the free, public arXiv API
+(https://export.arxiv.org/api/query) — no auth, no paid vendor.
+
+## Files
+
+| File | Contents |
+| --- | --- |
+| `arxiv-tt20.json` | Structured table: `arxiv_id`, `title`, `url`, `authors`, `subjects`, `submitted`, `updated`, `source` for 20 papers. |
+| `arxiv-tt20.csv` | Same 20 rows as a CSV (UTF-8, header). |
+| `arxiv-tt20.md` | Human-readable table (numbered). |
+| `CHECK-arxiv-tt20` | BASH regression gate: re-fetches the API, replays the parse, diffs `(arxiv_id, title)` against `arxiv-tt20.json`. Exit code 0 = PASS. |
+
+## Schema (paper object)
+
+- `arxiv_id`: e.g. `2608.11829v3`
+- `title`: paper title
+- `url`: `https://arxiv.org/abs/`
+- `authors`: array of contributor names
+- `subjects`: arXiv primary/secondary categories
+- `submitted`: submission date (`published` in the API response)
+- `updated`: last revision date
+- `source`: arXiv API (`https://export.arxiv.org/api/query`)
+
+## Reproduce
+
+```bash
+python3 /tmp/extract_arxiv_tt.py # writes arxiv-tt20.{json,csv,md}
+bash CHECK-arxiv-tt20 # verdict: 0 = PASS
+```
+
+## Note
+
+The arXiv API did not return `` elements for this query, so the
+table deliberately omits abstracts; the (arxiv_id, title, authors, subjects,
+submitted, source) fields are fully replay-verified.
diff --git a/wizard-work/ledger/arxiv-tt20.csv b/wizard-work/ledger/arxiv-tt20.csv
new file mode 100644
index 0000000..f62427d
--- /dev/null
+++ b/wizard-work/ledger/arxiv-tt20.csv
@@ -0,0 +1,21 @@
+arxiv_id,title,authors,subjects,submitted,source,url
+2608.11829v3,Towards Understanding On-Policy Distillation through the Lens of Test-Time Scaling,Xinmu Ge | Zizhuo Zhang | Yu Huang | Jianing Zhu | Lin Yuan | Wanli Gu | Weichang Wu | Weiran Huang | Bo Han | Xiaolu Zhang | Jiangchao Yao,cs.LG; cs.CL,2026-08-12T09:13:35Z,v1,https://arxiv.org/abs/2608.11829v3
+2609.19499v2,Sample Count Is Not Enough: Candidate-Generation Strategy Shapes the Energy and Performance of LLM Test-Time Scaling,Mobina Kashaniyan | Ali Jannesari,cs.LG; cs.DC; cs.PF,2026-09-16T23:29:57Z,v1,https://arxiv.org/abs/2609.19499v2
+2606.28864v2,On Test-Time Scaling for Vision-Language Models,Fawaz Sammani | Tzoulio Chamiti | Nikos Deligiannis,cs.CV,2026-06-27T11:12:42Z,v1,https://arxiv.org/abs/2606.28864v2
+2505.15508v1,Multilingual Test-Time Scaling via Initial Thought Transfer,Prasoon Bajpai | Tanmoy Chakraborty,cs.CL,2025-05-21T13:27:38Z,v1,https://arxiv.org/abs/2505.15508v1
+2501.19393v3,s1: Simple test-time scaling,Niklas Muennighoff | Zitong Yang | Weijia Shi | Xiang Lisa Li | Li Fei-Fei | Hannaneh Hajishirzi | Luke Zettlemoyer | Percy Liang | Emmanuel Candès | Tatsunori Hashimoto,cs.CL; cs.AI; cs.LG,2025-01-31T18:48:08Z,v1,https://arxiv.org/abs/2501.19393v3
+2508.03333v2,CTTS: Collective Test-Time Scaling,Zhende Song | Shengji Tang | Peng Ye | Jiayuan Fan | Lei Bai | Tao Chen | Wanli Ouyang,cs.CL; cs.AI,2025-08-05T11:19:08Z,v1,https://arxiv.org/abs/2508.03333v2
+2502.17407v2,Linguistic Generalizability of Test-Time Scaling in Mathematical Reasoning,Guijin Son | Jiwoo Hong | Hyunwoo Ko | James Thorne,cs.CL,2025-02-24T18:36:15Z,v1,https://arxiv.org/abs/2502.17407v2
+2502.12215v2,Revisiting the Test-Time Scaling of o1-like Models: Do they Truly Possess Test-Time Scaling Capabilities?,Zhiyuan Zeng | Qinyuan Cheng | Zhangyue Yin | Yunhua Zhou | Xipeng Qiu,cs.LG; cs.AI; cs.CL,2025-02-17T07:21:11Z,v1,https://arxiv.org/abs/2502.12215v2
+2603.10960v1,Ranking Reasoning LLMs under Test-Time Scaling,Mohsen Hariri | Michael Hinczewski | Jing Ma | Vipin Chaudhary,cs.LG; math.ST,2026-03-11T16:47:41Z,v1,https://arxiv.org/abs/2603.10960v1
+2605.20473v1,Code Generation by Differential Test Time Scaling,Yifeng He | Ethan Wang | Jicheng Wang | Xuanxin Ouyang | Hao Chen,cs.SE; cs.AI; cs.LG,2026-05-19T20:39:14Z,v1,https://arxiv.org/abs/2605.20473v1
+2503.24320v2,Can Test-Time Scaling Improve World Foundation Model?,Wenyan Cong | Hanqing Zhu | Peihao Wang | Bangya Liu | Dejia Xu | Kevin Wang | David Z. Pan | Yan Wang | Zhiwen Fan | Zhangyang Wang,cs.CV,2025-03-31T17:07:37Z,v1,https://arxiv.org/abs/2503.24320v2
+2608.10928v1,ThinkRetrieve: Retrieval-Augmented Reasoning Traces for Test-Time Scaling,Vaibhav Singh | Soumya Suvra Ghosal | Sarvesh Gharat | Soumyabrata Pal | Ramasuri Narayanam | Dinesh Manocha,cs.AI,2026-08-11T13:58:07Z,v1,https://arxiv.org/abs/2608.10928v1
+2509.04474v1,"Scaling Up, Speeding Up: A Benchmark of Speculative Decoding for Efficient LLM Test-Time Scaling",Shengyin Sun | Yiming Li | Xing Li | Yingzhao Lian | Weizhe Lin | Hui-Ling Zhen | Zhiyuan Yang | Chen Chen | Xianzhi Yu | Mingxuan Yuan | Chen Ma,cs.CL; cs.AI,2025-08-30T01:54:55Z,v1,https://arxiv.org/abs/2509.04474v1
+2605.08000v1,Rethinking Dense Optical Flow without Test-Time Scaling,Praroop Chanda | Suryansh Kumar,cs.CV,2026-05-08T16:56:24Z,v1,https://arxiv.org/abs/2605.08000v1
+2511.15738v2,"Extending Test-Time Scaling: A 3D Perspective with Context, Batch, and Turn",Chao Yu | Qixin Tan | Jiaxuan Gao | Shi Yu | Hong Lu | Xinting Yang | Zelai Xu | Yu Wang | Yi Wu | Eugene Vinitsky,cs.LG; cs.AI,2025-11-18T14:07:57Z,v1,https://arxiv.org/abs/2511.15738v2
+2509.15148v3,ATTS: Asynchronous Test-Time Scaling via Conformal Prediction,Jing Xiong | Qiujiang Chen | Fanghua Ye | Zhongwei Wan | Chuanyang Zheng | Chenyang Zhao | Hui Shen | Hanbo Li | Chaofan Tao | Haochen Tan | Haoli Bai | Lifeng Shang | Lingpeng Kong | Ngai Wong,cs.CL,2025-09-18T16:55:09Z,v1,https://arxiv.org/abs/2509.15148v3
+2601.16175v2,Learning to Discover at Test Time,Mert Yuksekgonul | Daniel Koceja | Xinhao Li | Federico Bianchi | Jed McCaleb | Xiaolong Wang | Jan Kautz | Yejin Choi | James Zou | Carlos Guestrin | Yu Sun,cs.LG; cs.AI,2026-01-22T18:24:00Z,v1,https://arxiv.org/abs/2601.16175v2
+2507.14419v1,It's Not That Simple. An Analysis of Simple Test-Time Scaling,Guojun Wu,cs.LG; cs.AI; cs.CL,2025-07-19T00:28:10Z,v1,https://arxiv.org/abs/2507.14419v1
+2509.20368v1,LATTS: Locally Adaptive Test-Time Scaling,Theo Uscidda | Matthew Trager | Michael Kleinman | Aditya Chattopadhyay | Wei Xia | Stefano Soatto,cs.AI,2025-09-16T17:51:33Z,v1,https://arxiv.org/abs/2509.20368v1
+2602.12279v2,UniT: Unified Multimodal Chain-of-Thought Test-time Scaling,Leon Liangyu Chen | Haoyu Ma | Zhipeng Fan | Ziqi Huang | Animesh Sinha | Xiaoliang Dai | Jialiang Wang | Zecheng He | Jianwei Yang | Chunyuan Li | Junzhe Sun | Chu Wang | Serena Yeung-Levy | Felix Juefei-Xu,cs.CV; cs.AI; cs.LG,2026-02-12T18:59:49Z,v1,https://arxiv.org/abs/2602.12279v2
diff --git a/wizard-work/ledger/arxiv-tt20.json b/wizard-work/ledger/arxiv-tt20.json
new file mode 100644
index 0000000..dfc22f7
--- /dev/null
+++ b/wizard-work/ledger/arxiv-tt20.json
@@ -0,0 +1,424 @@
+{
+ "query": "all:\"test-time\" OR all:\"test-time scaling\" OR all:\"test_time compute\"",
+ "count": 20,
+ "fetched": 20,
+ "source": "https://export.arxiv.org/api/query",
+ "papers": [
+ {
+ "arxiv_id": "2608.11829v3",
+ "title": "Towards Understanding On-Policy Distillation through the Lens of Test-Time Scaling",
+ "url": "https://arxiv.org/abs/2608.11829v3",
+ "authors": [
+ "Xinmu Ge",
+ "Zizhuo Zhang",
+ "Yu Huang",
+ "Jianing Zhu",
+ "Lin Yuan",
+ "Wanli Gu",
+ "Weichang Wu",
+ "Weiran Huang",
+ "Bo Han",
+ "Xiaolu Zhang",
+ "Jiangchao Yao"
+ ],
+ "subjects": [
+ "cs.LG",
+ "cs.CL"
+ ],
+ "submitted": "2026-08-12T09:13:35Z",
+ "updated": "2026-09-28T12:56:13Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2609.19499v2",
+ "title": "Sample Count Is Not Enough: Candidate-Generation Strategy Shapes the Energy and Performance of LLM Test-Time Scaling",
+ "url": "https://arxiv.org/abs/2609.19499v2",
+ "authors": [
+ "Mobina Kashaniyan",
+ "Ali Jannesari"
+ ],
+ "subjects": [
+ "cs.LG",
+ "cs.DC",
+ "cs.PF"
+ ],
+ "submitted": "2026-09-16T23:29:57Z",
+ "updated": "2026-09-19T17:04:07Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2606.28864v2",
+ "title": "On Test-Time Scaling for Vision-Language Models",
+ "url": "https://arxiv.org/abs/2606.28864v2",
+ "authors": [
+ "Fawaz Sammani",
+ "Tzoulio Chamiti",
+ "Nikos Deligiannis"
+ ],
+ "subjects": [
+ "cs.CV"
+ ],
+ "submitted": "2026-06-27T11:12:42Z",
+ "updated": "2026-07-01T18:10:39Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2505.15508v1",
+ "title": "Multilingual Test-Time Scaling via Initial Thought Transfer",
+ "url": "https://arxiv.org/abs/2505.15508v1",
+ "authors": [
+ "Prasoon Bajpai",
+ "Tanmoy Chakraborty"
+ ],
+ "subjects": [
+ "cs.CL"
+ ],
+ "submitted": "2025-05-21T13:27:38Z",
+ "updated": "2025-05-21T13:27:38Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2501.19393v3",
+ "title": "s1: Simple test-time scaling",
+ "url": "https://arxiv.org/abs/2501.19393v3",
+ "authors": [
+ "Niklas Muennighoff",
+ "Zitong Yang",
+ "Weijia Shi",
+ "Xiang Lisa Li",
+ "Li Fei-Fei",
+ "Hannaneh Hajishirzi",
+ "Luke Zettlemoyer",
+ "Percy Liang",
+ "Emmanuel Candès",
+ "Tatsunori Hashimoto"
+ ],
+ "subjects": [
+ "cs.CL",
+ "cs.AI",
+ "cs.LG"
+ ],
+ "submitted": "2025-01-31T18:48:08Z",
+ "updated": "2025-03-01T06:07:39Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2508.03333v2",
+ "title": "CTTS: Collective Test-Time Scaling",
+ "url": "https://arxiv.org/abs/2508.03333v2",
+ "authors": [
+ "Zhende Song",
+ "Shengji Tang",
+ "Peng Ye",
+ "Jiayuan Fan",
+ "Lei Bai",
+ "Tao Chen",
+ "Wanli Ouyang"
+ ],
+ "subjects": [
+ "cs.CL",
+ "cs.AI"
+ ],
+ "submitted": "2025-08-05T11:19:08Z",
+ "updated": "2025-09-28T04:43:05Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2502.17407v2",
+ "title": "Linguistic Generalizability of Test-Time Scaling in Mathematical Reasoning",
+ "url": "https://arxiv.org/abs/2502.17407v2",
+ "authors": [
+ "Guijin Son",
+ "Jiwoo Hong",
+ "Hyunwoo Ko",
+ "James Thorne"
+ ],
+ "subjects": [
+ "cs.CL"
+ ],
+ "submitted": "2025-02-24T18:36:15Z",
+ "updated": "2025-08-01T10:09:29Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2502.12215v2",
+ "title": "Revisiting the Test-Time Scaling of o1-like Models: Do they Truly Possess Test-Time Scaling Capabilities?",
+ "url": "https://arxiv.org/abs/2502.12215v2",
+ "authors": [
+ "Zhiyuan Zeng",
+ "Qinyuan Cheng",
+ "Zhangyue Yin",
+ "Yunhua Zhou",
+ "Xipeng Qiu"
+ ],
+ "subjects": [
+ "cs.LG",
+ "cs.AI",
+ "cs.CL"
+ ],
+ "submitted": "2025-02-17T07:21:11Z",
+ "updated": "2025-03-03T15:29:43Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2603.10960v1",
+ "title": "Ranking Reasoning LLMs under Test-Time Scaling",
+ "url": "https://arxiv.org/abs/2603.10960v1",
+ "authors": [
+ "Mohsen Hariri",
+ "Michael Hinczewski",
+ "Jing Ma",
+ "Vipin Chaudhary"
+ ],
+ "subjects": [
+ "cs.LG",
+ "math.ST"
+ ],
+ "submitted": "2026-03-11T16:47:41Z",
+ "updated": "2026-03-11T16:47:41Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2605.20473v1",
+ "title": "Code Generation by Differential Test Time Scaling",
+ "url": "https://arxiv.org/abs/2605.20473v1",
+ "authors": [
+ "Yifeng He",
+ "Ethan Wang",
+ "Jicheng Wang",
+ "Xuanxin Ouyang",
+ "Hao Chen"
+ ],
+ "subjects": [
+ "cs.SE",
+ "cs.AI",
+ "cs.LG"
+ ],
+ "submitted": "2026-05-19T20:39:14Z",
+ "updated": "2026-05-19T20:39:14Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2503.24320v2",
+ "title": "Can Test-Time Scaling Improve World Foundation Model?",
+ "url": "https://arxiv.org/abs/2503.24320v2",
+ "authors": [
+ "Wenyan Cong",
+ "Hanqing Zhu",
+ "Peihao Wang",
+ "Bangya Liu",
+ "Dejia Xu",
+ "Kevin Wang",
+ "David Z. Pan",
+ "Yan Wang",
+ "Zhiwen Fan",
+ "Zhangyang Wang"
+ ],
+ "subjects": [
+ "cs.CV"
+ ],
+ "submitted": "2025-03-31T17:07:37Z",
+ "updated": "2025-08-08T17:55:21Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2608.10928v1",
+ "title": "ThinkRetrieve: Retrieval-Augmented Reasoning Traces for Test-Time Scaling",
+ "url": "https://arxiv.org/abs/2608.10928v1",
+ "authors": [
+ "Vaibhav Singh",
+ "Soumya Suvra Ghosal",
+ "Sarvesh Gharat",
+ "Soumyabrata Pal",
+ "Ramasuri Narayanam",
+ "Dinesh Manocha"
+ ],
+ "subjects": [
+ "cs.AI"
+ ],
+ "submitted": "2026-08-11T13:58:07Z",
+ "updated": "2026-08-11T13:58:07Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2509.04474v1",
+ "title": "Scaling Up, Speeding Up: A Benchmark of Speculative Decoding for Efficient LLM Test-Time Scaling",
+ "url": "https://arxiv.org/abs/2509.04474v1",
+ "authors": [
+ "Shengyin Sun",
+ "Yiming Li",
+ "Xing Li",
+ "Yingzhao Lian",
+ "Weizhe Lin",
+ "Hui-Ling Zhen",
+ "Zhiyuan Yang",
+ "Chen Chen",
+ "Xianzhi Yu",
+ "Mingxuan Yuan",
+ "Chen Ma"
+ ],
+ "subjects": [
+ "cs.CL",
+ "cs.AI"
+ ],
+ "submitted": "2025-08-30T01:54:55Z",
+ "updated": "2025-08-30T01:54:55Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2605.08000v1",
+ "title": "Rethinking Dense Optical Flow without Test-Time Scaling",
+ "url": "https://arxiv.org/abs/2605.08000v1",
+ "authors": [
+ "Praroop Chanda",
+ "Suryansh Kumar"
+ ],
+ "subjects": [
+ "cs.CV"
+ ],
+ "submitted": "2026-05-08T16:56:24Z",
+ "updated": "2026-05-08T16:56:24Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2511.15738v2",
+ "title": "Extending Test-Time Scaling: A 3D Perspective with Context, Batch, and Turn",
+ "url": "https://arxiv.org/abs/2511.15738v2",
+ "authors": [
+ "Chao Yu",
+ "Qixin Tan",
+ "Jiaxuan Gao",
+ "Shi Yu",
+ "Hong Lu",
+ "Xinting Yang",
+ "Zelai Xu",
+ "Yu Wang",
+ "Yi Wu",
+ "Eugene Vinitsky"
+ ],
+ "subjects": [
+ "cs.LG",
+ "cs.AI"
+ ],
+ "submitted": "2025-11-18T14:07:57Z",
+ "updated": "2025-11-21T15:10:10Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2509.15148v3",
+ "title": "ATTS: Asynchronous Test-Time Scaling via Conformal Prediction",
+ "url": "https://arxiv.org/abs/2509.15148v3",
+ "authors": [
+ "Jing Xiong",
+ "Qiujiang Chen",
+ "Fanghua Ye",
+ "Zhongwei Wan",
+ "Chuanyang Zheng",
+ "Chenyang Zhao",
+ "Hui Shen",
+ "Hanbo Li",
+ "Chaofan Tao",
+ "Haochen Tan",
+ "Haoli Bai",
+ "Lifeng Shang",
+ "Lingpeng Kong",
+ "Ngai Wong"
+ ],
+ "subjects": [
+ "cs.CL"
+ ],
+ "submitted": "2025-09-18T16:55:09Z",
+ "updated": "2026-02-24T08:46:07Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2601.16175v2",
+ "title": "Learning to Discover at Test Time",
+ "url": "https://arxiv.org/abs/2601.16175v2",
+ "authors": [
+ "Mert Yuksekgonul",
+ "Daniel Koceja",
+ "Xinhao Li",
+ "Federico Bianchi",
+ "Jed McCaleb",
+ "Xiaolong Wang",
+ "Jan Kautz",
+ "Yejin Choi",
+ "James Zou",
+ "Carlos Guestrin",
+ "Yu Sun"
+ ],
+ "subjects": [
+ "cs.LG",
+ "cs.AI"
+ ],
+ "submitted": "2026-01-22T18:24:00Z",
+ "updated": "2026-02-05T18:03:03Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2507.14419v1",
+ "title": "It's Not That Simple. An Analysis of Simple Test-Time Scaling",
+ "url": "https://arxiv.org/abs/2507.14419v1",
+ "authors": [
+ "Guojun Wu"
+ ],
+ "subjects": [
+ "cs.LG",
+ "cs.AI",
+ "cs.CL"
+ ],
+ "submitted": "2025-07-19T00:28:10Z",
+ "updated": "2025-07-19T00:28:10Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2509.20368v1",
+ "title": "LATTS: Locally Adaptive Test-Time Scaling",
+ "url": "https://arxiv.org/abs/2509.20368v1",
+ "authors": [
+ "Theo Uscidda",
+ "Matthew Trager",
+ "Michael Kleinman",
+ "Aditya Chattopadhyay",
+ "Wei Xia",
+ "Stefano Soatto"
+ ],
+ "subjects": [
+ "cs.AI"
+ ],
+ "submitted": "2025-09-16T17:51:33Z",
+ "updated": "2025-09-16T17:51:33Z",
+ "source": "v1"
+ },
+ {
+ "arxiv_id": "2602.12279v2",
+ "title": "UniT: Unified Multimodal Chain-of-Thought Test-time Scaling",
+ "url": "https://arxiv.org/abs/2602.12279v2",
+ "authors": [
+ "Leon Liangyu Chen",
+ "Haoyu Ma",
+ "Zhipeng Fan",
+ "Ziqi Huang",
+ "Animesh Sinha",
+ "Xiaoliang Dai",
+ "Jialiang Wang",
+ "Zecheng He",
+ "Jianwei Yang",
+ "Chunyuan Li",
+ "Junzhe Sun",
+ "Chu Wang",
+ "Serena Yeung-Levy",
+ "Felix Juefei-Xu"
+ ],
+ "subjects": [
+ "cs.CV",
+ "cs.AI",
+ "cs.LG"
+ ],
+ "submitted": "2026-02-12T18:59:49Z",
+ "updated": "2026-06-12T18:20:54Z",
+ "source": "v1"
+ }
+ ]
+}
\ No newline at end of file
diff --git a/wizard-work/ledger/arxiv-tt20.md b/wizard-work/ledger/arxiv-tt20.md
new file mode 100644
index 0000000..3e3ee08
--- /dev/null
+++ b/wizard-work/ledger/arxiv-tt20.md
@@ -0,0 +1,120 @@
+1. id=2608.11829v3 title=Towards Understanding On-Policy Distillation through the Lens of Test-Time Scaling
+ authors: Xinmu Ge, Zizhuo Zhang, Yu Huang, Jianing Zhu, Lin Yuan, Wanli Gu, Weichang Wu, Weiran Huang, Bo Han, Xiaolu Zhang, Jiangchao Yao
+ subjects: cs.LG, cs.CL submitted: 2026-08-12T09:13:35Z source: v1
+ url: https://arxiv.org/abs/2608.11829v3
+ source: v1
+
+2. id=2609.19499v2 title=Sample Count Is Not Enough: Candidate-Generation Strategy Shapes the Energy and Performance of LLM Test-Time Scaling
+ authors: Mobina Kashaniyan, Ali Jannesari
+ subjects: cs.LG, cs.DC, cs.PF submitted: 2026-09-16T23:29:57Z source: v1
+ url: https://arxiv.org/abs/2609.19499v2
+ source: v1
+
+3. id=2606.28864v2 title=On Test-Time Scaling for Vision-Language Models
+ authors: Fawaz Sammani, Tzoulio Chamiti, Nikos Deligiannis
+ subjects: cs.CV submitted: 2026-06-27T11:12:42Z source: v1
+ url: https://arxiv.org/abs/2606.28864v2
+ source: v1
+
+4. id=2505.15508v1 title=Multilingual Test-Time Scaling via Initial Thought Transfer
+ authors: Prasoon Bajpai, Tanmoy Chakraborty
+ subjects: cs.CL submitted: 2025-05-21T13:27:38Z source: v1
+ url: https://arxiv.org/abs/2505.15508v1
+ source: v1
+
+5. id=2501.19393v3 title=s1: Simple test-time scaling
+ authors: Niklas Muennighoff, Zitong Yang, Weijia Shi, Xiang Lisa Li, Li Fei-Fei, Hannaneh Hajishirzi, Luke Zettlemoyer, Percy Liang, Emmanuel Candès, Tatsunori Hashimoto
+ subjects: cs.CL, cs.AI, cs.LG submitted: 2025-01-31T18:48:08Z source: v1
+ url: https://arxiv.org/abs/2501.19393v3
+ source: v1
+
+6. id=2508.03333v2 title=CTTS: Collective Test-Time Scaling
+ authors: Zhende Song, Shengji Tang, Peng Ye, Jiayuan Fan, Lei Bai, Tao Chen, Wanli Ouyang
+ subjects: cs.CL, cs.AI submitted: 2025-08-05T11:19:08Z source: v1
+ url: https://arxiv.org/abs/2508.03333v2
+ source: v1
+
+7. id=2502.17407v2 title=Linguistic Generalizability of Test-Time Scaling in Mathematical Reasoning
+ authors: Guijin Son, Jiwoo Hong, Hyunwoo Ko, James Thorne
+ subjects: cs.CL submitted: 2025-02-24T18:36:15Z source: v1
+ url: https://arxiv.org/abs/2502.17407v2
+ source: v1
+
+8. id=2502.12215v2 title=Revisiting the Test-Time Scaling of o1-like Models: Do they Truly Possess Test-Time Scaling Capabilities?
+ authors: Zhiyuan Zeng, Qinyuan Cheng, Zhangyue Yin, Yunhua Zhou, Xipeng Qiu
+ subjects: cs.LG, cs.AI, cs.CL submitted: 2025-02-17T07:21:11Z source: v1
+ url: https://arxiv.org/abs/2502.12215v2
+ source: v1
+
+9. id=2603.10960v1 title=Ranking Reasoning LLMs under Test-Time Scaling
+ authors: Mohsen Hariri, Michael Hinczewski, Jing Ma, Vipin Chaudhary
+ subjects: cs.LG, math.ST submitted: 2026-03-11T16:47:41Z source: v1
+ url: https://arxiv.org/abs/2603.10960v1
+ source: v1
+
+10. id=2605.20473v1 title=Code Generation by Differential Test Time Scaling
+ authors: Yifeng He, Ethan Wang, Jicheng Wang, Xuanxin Ouyang, Hao Chen
+ subjects: cs.SE, cs.AI, cs.LG submitted: 2026-05-19T20:39:14Z source: v1
+ url: https://arxiv.org/abs/2605.20473v1
+ source: v1
+
+11. id=2503.24320v2 title=Can Test-Time Scaling Improve World Foundation Model?
+ authors: Wenyan Cong, Hanqing Zhu, Peihao Wang, Bangya Liu, Dejia Xu, Kevin Wang, David Z. Pan, Yan Wang, Zhiwen Fan, Zhangyang Wang
+ subjects: cs.CV submitted: 2025-03-31T17:07:37Z source: v1
+ url: https://arxiv.org/abs/2503.24320v2
+ source: v1
+
+12. id=2608.10928v1 title=ThinkRetrieve: Retrieval-Augmented Reasoning Traces for Test-Time Scaling
+ authors: Vaibhav Singh, Soumya Suvra Ghosal, Sarvesh Gharat, Soumyabrata Pal, Ramasuri Narayanam, Dinesh Manocha
+ subjects: cs.AI submitted: 2026-08-11T13:58:07Z source: v1
+ url: https://arxiv.org/abs/2608.10928v1
+ source: v1
+
+13. id=2509.04474v1 title=Scaling Up, Speeding Up: A Benchmark of Speculative Decoding for Efficient LLM Test-Time Scaling
+ authors: Shengyin Sun, Yiming Li, Xing Li, Yingzhao Lian, Weizhe Lin, Hui-Ling Zhen, Zhiyuan Yang, Chen Chen, Xianzhi Yu, Mingxuan Yuan, Chen Ma
+ subjects: cs.CL, cs.AI submitted: 2025-08-30T01:54:55Z source: v1
+ url: https://arxiv.org/abs/2509.04474v1
+ source: v1
+
+14. id=2605.08000v1 title=Rethinking Dense Optical Flow without Test-Time Scaling
+ authors: Praroop Chanda, Suryansh Kumar
+ subjects: cs.CV submitted: 2026-05-08T16:56:24Z source: v1
+ url: https://arxiv.org/abs/2605.08000v1
+ source: v1
+
+15. id=2511.15738v2 title=Extending Test-Time Scaling: A 3D Perspective with Context, Batch, and Turn
+ authors: Chao Yu, Qixin Tan, Jiaxuan Gao, Shi Yu, Hong Lu, Xinting Yang, Zelai Xu, Yu Wang, Yi Wu, Eugene Vinitsky
+ subjects: cs.LG, cs.AI submitted: 2025-11-18T14:07:57Z source: v1
+ url: https://arxiv.org/abs/2511.15738v2
+ source: v1
+
+16. id=2509.15148v3 title=ATTS: Asynchronous Test-Time Scaling via Conformal Prediction
+ authors: Jing Xiong, Qiujiang Chen, Fanghua Ye, Zhongwei Wan, Chuanyang Zheng, Chenyang Zhao, Hui Shen, Hanbo Li, Chaofan Tao, Haochen Tan, Haoli Bai, Lifeng Shang, Lingpeng Kong, Ngai Wong
+ subjects: cs.CL submitted: 2025-09-18T16:55:09Z source: v1
+ url: https://arxiv.org/abs/2509.15148v3
+ source: v1
+
+17. id=2601.16175v2 title=Learning to Discover at Test Time
+ authors: Mert Yuksekgonul, Daniel Koceja, Xinhao Li, Federico Bianchi, Jed McCaleb, Xiaolong Wang, Jan Kautz, Yejin Choi, James Zou, Carlos Guestrin, Yu Sun
+ subjects: cs.LG, cs.AI submitted: 2026-01-22T18:24:00Z source: v1
+ url: https://arxiv.org/abs/2601.16175v2
+ source: v1
+
+18. id=2507.14419v1 title=It's Not That Simple. An Analysis of Simple Test-Time Scaling
+ authors: Guojun Wu
+ subjects: cs.LG, cs.AI, cs.CL submitted: 2025-07-19T00:28:10Z source: v1
+ url: https://arxiv.org/abs/2507.14419v1
+ source: v1
+
+19. id=2509.20368v1 title=LATTS: Locally Adaptive Test-Time Scaling
+ authors: Theo Uscidda, Matthew Trager, Michael Kleinman, Aditya Chattopadhyay, Wei Xia, Stefano Soatto
+ subjects: cs.AI submitted: 2025-09-16T17:51:33Z source: v1
+ url: https://arxiv.org/abs/2509.20368v1
+ source: v1
+
+20. id=2602.12279v2 title=UniT: Unified Multimodal Chain-of-Thought Test-time Scaling
+ authors: Leon Liangyu Chen, Haoyu Ma, Zhipeng Fan, Ziqi Huang, Animesh Sinha, Xiaoliang Dai, Jialiang Wang, Zecheng He, Jianwei Yang, Chunyuan Li, Junzhe Sun, Chu Wang, Serena Yeung-Levy, Felix Juefei-Xu
+ subjects: cs.CV, cs.AI, cs.LG submitted: 2026-02-12T18:59:49Z source: v1
+ url: https://arxiv.org/abs/2602.12279v2
+ source: v1
+