pith:FCKZ6JJK
Quantifying and Mitigating Self-Preference Bias of LLM Judges
LLM judges show self-preference bias uncorrelated with capability, but a multi-dimensional strategy reduces it by 31.5 percent.
arxiv:2604.22891 v4 · 2026-04-24 · cs.LG · cs.AI · cs.CL
Add to your LaTeX paper
\usepackage{pith}
\pithnumber{FCKZ6JJKVWJM35PNRHLWIGEFAZ}
Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge
Record completeness
Claims
Empirical analysis across 20 mainstream LLMs reveals that advanced capabilities are often uncorrelated, or even negatively correlated, with low SPB. To mitigate this bias, we propose a structured multi-dimensional evaluation strategy grounded in cognitive load decomposition, which reduces SPB by 31.5% on average.
The constructed pairs of responses truly have negligible quality differences, allowing statistical separation of bias propensity from genuine discriminability without human gold standards.
An automated framework using equal-quality response pairs quantifies self-preference bias in LLM judges and reduces it by 31.5% via a cognitive-load-based multi-dimensional evaluation strategy.
Formal links
Receipt and verification
| First computed | 2026-06-03T01:05:50.628659Z |
|---|---|
| Builder | pith-number-builder-2026-05-17-v1 |
| Signature | Pith Ed25519
(pith-v1-2026-05) · public key |
| Schema | pith-number/v1.0 |
Canonical hash
28959f252aad92cdf5ed89d76418850669e7c6d48051135a8709401c7fa1e9f0
Aliases
· · · · ·Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/FCKZ6JJKVWJM35PNRHLWIGEFAZ \
| 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: 28959f252aad92cdf5ed89d76418850669e7c6d48051135a8709401c7fa1e9f0
Canonical record JSON
{
"metadata": {
"abstract_canon_sha256": "0e61c9836c571c203405ee11af53a5bc9430944ca9704f8f9f98e09b091f3d38",
"cross_cats_sorted": [
"cs.AI",
"cs.CL"
],
"license": "http://creativecommons.org/licenses/by/4.0/",
"primary_cat": "cs.LG",
"submitted_at": "2026-04-24T09:46:22Z",
"title_canon_sha256": "4d4d23e09121506db1381c0fccf13bf15848e72874277c48229f8414b8ea0d8d"
},
"schema_version": "1.0",
"source": {
"id": "2604.22891",
"kind": "arxiv",
"version": 4
}
}