DPO (Mixture-of-experts routing): superseded — cited as a baseline and beaten by newer methods. 2 paper(s) critique it, 1 beat it on benchmarks — #54 of 1370 most-superseded. Sub-problem: cluster led by SteerMoE. Newer alternatives in the same sub-problem include PADD, MESA, PR2, RouteHijack, MASCing.

Superseded baseline#54 of 1,370 most-superseded

DPO

Mixture-of-experts routing

superseded — cited as a baseline and beaten by newer methods

2 papers critique it · 1 beat it on benchmarks

What papers say

Verbatim critique sentences, each from a paper that cites DPO as a baseline.

“requires explicit human or teacher model judgments to label one response as preferred over another, making them costly and hard to scale”
— OrdMoE: Preference Alignment via Hierarchical Expert Group Ranking in Multimodal Mixture-of-Experts LLMs
“standard DPO methods are still inherently limited by their reliance on a single, monolithic policy.”
— Mix- and MoE-DPO: A Variational Inference Approach to Direct Preference Optimization

Beaten on benchmarks

Head-to-head results where a newer method reports beating DPO. Values are copied from the source paper's tables — verify against the cited paper.

GSPO+SSR (Ours) beats DPO · BERTScore-F1 [no condition]
0.506 vs 0.472
SocialNav-MoE: A Mixture-of-Experts Vision Language Model for Socially Compliant Navigation with Reinforcement Fine-Tuning
GSPO+SSR (Ours) beats DPO · SBERT-cos [no condition]
0.671 vs 0.649
SocialNav-MoE: A Mixture-of-Experts Vision Language Model for Socially Compliant Navigation with Reinforcement Fine-Tuning
GSPO+SSR (Ours) beats DPO · SMS [no condition]
0.551 vs 0.527
SocialNav-MoE: A Mixture-of-Experts Vision Language Model for Socially Compliant Navigation with Reinforcement Fine-Tuning

Newer alternatives

Recent methods in the same sub-problem, not yet superseded in the knowledge base.