Metraipolis › archive › dc_f0e7ef6486f4

Swapping the Judge Changes the Ranking: A Preregistered Study of Cross-Provider Replication in LLM-Judged System Comparisons — arXiv:2611.04090

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.04090
License (as archived)
CC0-1.0
Archived
2026-11-04T18:36:26.427Z
SHA-256 of the archived text
6352f59fc4ba4dfcbbb76d8b5ed43d91a99bec776a44d216b541ad5614eb3663
Stable URL of this page
https://metraipolis.com/documento/dc_f0e7ef6486f4

How to cite

Swapping the Judge Changes the Ranking: A Preregistered Study of Cross-Provider Replication in LLM-Judged System Comparisons — arXiv:2611.04090 — arXiv (en), https://arxiv.org/abs/2611.04090 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f0e7ef6486f4 (Metraipolis archive dc_f0e7ef6486f4, sha256 6352f59fc4ba4dfcbbb76d8b5ed43d91a99bec776a44d216b541ad5614eb3663)
BibTeX
@misc{dcf0e7ef6486f4,
  title = {Swapping the Judge Changes the Ranking: A Preregistered Study of Cross-Provider Replication in LLM-Judged System Comparisons — arXiv:2611.04090},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.04090},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f0e7ef6486f4; sha256 6352f59fc4ba4dfcbbb76d8b5ed43d91a99bec776a44d216b541ad5614eb3663},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.04090. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Practitioners who compare systems with an LLM judge and have no human labels often check a conclusion by repeating it with a judge from another provider. We test how far that check can be trusted. In a preregistered study (with a disclosed post-registration reduction of the judge set), five judge models from three providers re-judged 4,000 Chatbot Arena battles (nine systems, 33 system pairs) that carry human votes, each in both response orders. Agreement with the human vote was low at the verdict level (median Cohen's kappa = 0.22), and system rankings did not survive a change of judge: Kendall's tau with the human ranking ranged from 0.17 to 0.89 (median 0.67), so the preregistered hypothesis that rankings remain stable was not supported. Conclusions replicated by two low-cost judges from different providers matched the human conclusion more often than single-judge conclusions (+0.10 precision), but the gain came from filtering out one poorly aligned judge, and replication involving that judge left a conclusion for only 35% of comparisons. Judges that agreed with humans reached a stable conclusion with no more verdicts than human votes required; the others failed to converge on 26-70% of system pairs at the sample sizes tested (up to 100 verdicts per pair). Disagreements with the human vote co-occurred across judges far more often than independence predicts, although a single human vote per battle cannot separate shared judge bias from voter noise.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI); Computers and Society (cs.CY).
Pagina del lavoro: https://arxiv.org/abs/2611.04090 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-04T18:32:37.413Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).