Pith Number

pith:JPFVERVP

pith:2026:JPFVERVPP227NRW7IOZEYKWBYT

not attested not anchored not stored refs resolved

Effective Harness Engineering for Algorithm Discovery with Coding Agents

Masafumi Oyamada, Taro Yano, Yoichi Ishibashi

Generating fewer algorithms with deeper thought outperforms many brief ones under a fixed token budget in algorithm discovery.

arxiv:2605.15221 v1 · 2026-05-13 · cs.SE · cs.AI · cs.CL

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{JPFVERVPP227NRW7IOZEYKWBYT}

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

Generating fewer algorithms while thinking more deeply about each one achieved higher scores. That is, scaling the quality of each individual is more budget-efficient than scaling the number of evolutionary generations.

C2weakest assumption

That results observed on the single Circle Packing benchmark under one fixed token budget generalize to other algorithm-discovery tasks and different resource constraints.

C3one line summary

Under fixed token budget on Circle Packing, deeper per-candidate reasoning beats generating more shallow candidates, and capable models produce evaluation hacks at higher rates.

References

26 extracted · 26 resolved · 5 Pith anchors

[1] AlphaEvolve: A coding agent for scientific and algorithmic discovery 2025 · arXiv:2506.13131

[2] Mathematical discoveries from program search with large language models , journal = 2024

[3] arXiv preprint arXiv:2206.08896 , year= 2022

[4] Forty-first International Conference on Machine Learning, 2024

[5] Haoran Ye and Jiarui Wang and Zhiguang Cao and Federico Berto and Chuanbo Hua and Haeyeon Kim and Jinkyoo Park and Guojie Song , title =. Advances in Neural Information Processing Systems 38: Annual C 2024

Receipt and verification

First computed	2026-05-20T00:00:46.968684Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

4bcb5246af7eb5f6c6df43b24c2ac1c4edfc8a657be360285890297ecaa8abee

Aliases

arxiv: 2605.15221 · arxiv_version: 2605.15221v1 · doi: 10.48550/arxiv.2605.15221 · pith_short_12: JPFVERVPP227 · pith_short_16: JPFVERVPP227NRW7 · pith_short_8: JPFVERVP

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/JPFVERVPP227NRW7IOZEYKWBYT \
  | 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: 4bcb5246af7eb5f6c6df43b24c2ac1c4edfc8a657be360285890297ecaa8abee

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "b5319c43efe5140779b2ac4cfed0b33089f138b114c12753bd962c47f17e5909",
    "cross_cats_sorted": [
      "cs.AI",
      "cs.CL"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.SE",
    "submitted_at": "2026-05-13T06:33:06Z",
    "title_canon_sha256": "405e377103930fc743f407ef7936ac8bbbd034df49937660257abb639beda96b"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.15221",
    "kind": "arxiv",
    "version": 1
  }
}