Pith Number

pith:QZEGLJK3

pith:2026:QZEGLJK3QZ66X3XP2R7FJSJGRS

not attested not anchored not stored refs resolved

CAPS: Cascaded Adaptive Pairwise Selection for Efficient Parallel Reasoning

Fangzhou Lin, Haichong Zhang, Kazunori Yamada, Peiran Li, Qianwen Ge, Shuo Xing, Siyuan Yang, Zhengzhong Tu, Ziming Zhang

CAPS uses a four-stage cascade to adapt evidence and pair selection so pairwise verification costs far less while selecting better answers.

arxiv:2605.15513 v1 · 2026-05-15 · cs.AI

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{QZEGLJK3QZ66X3XP2R7FJSJGRS}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

On four self-verifying models and five reasoning benchmarks, CAPS outperforms the leading pairwise verifier on 14 of 20 suites while using 25.4% of its verifier-token budget on code, and outperforms pointwise self-verification on all 20.

C2weakest assumption

The verifier's accuracy at partial versus full evidence remains high enough that early-stage decisions in the cascade do not introduce irrecoverable errors; this is invoked in the description of the four-stage cascade and the optional rescue subroutine.

C3one line summary

CAPS is a four-stage inference-only cascade that adapts how much of each solution the verifier sees and how comparisons are distributed, halving per-candidate verifier tokens while outperforming uniform pairwise verification on most benchmarks.

References

65 extracted · 65 resolved · 13 Pith anchors

[1] gpt-oss-120b & gpt-oss-20b Model Card 2025 · arXiv:2508.10925

[2] MathArena: Evaluating LLMs on Uncontaminated Math Competitions 2025 · arXiv:2505.23281

[3] Large Language Monkeys: Scaling Inference Compute with Repeated Sampling 2024 · arXiv:2407.21787

[4] Evaluating Large Language Models Trained on Code 2021 · arXiv:2107.03374

[5] Universal self-consistency for large language model generation 2023

Formal links

1 machine-checked theorem link

Cited by

1 paper in Pith

PathCal: State-Aware Reflection-Marker Calibration for Efficient Reasoning

Receipt and verification

First computed	2026-05-20T00:01:02.525081Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

864865a55b867debeeefd47e54c9268c8ce8b74cc647b64978e1626506c02943

Aliases

arxiv: 2605.15513 · arxiv_version: 2605.15513v1 · doi: 10.48550/arxiv.2605.15513 · pith_short_12: QZEGLJK3QZ66 · pith_short_16: QZEGLJK3QZ66X3XP · pith_short_8: QZEGLJK3

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/QZEGLJK3QZ66X3XP2R7FJSJGRS \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: 864865a55b867debeeefd47e54c9268c8ce8b74cc647b64978e1626506c02943

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "6ace4a5c8d287c180dc64f30324a82de6ef52b94149ced58c899fe87c0ef5975",
    "cross_cats_sorted": [],
    "license": "http://creativecommons.org/licenses/by-nc-sa/4.0/",
    "primary_cat": "cs.AI",
    "submitted_at": "2026-05-15T01:16:12Z",
    "title_canon_sha256": "54277e13d70bd6b643130666afa2e5fd9126a0145533b2ddd8d918e440a12a3b"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.15513",
    "kind": "arxiv",
    "version": 1
  }
}