Query-aware routing for Cross-lingual performance gains in Encoders

Akshay Jain; Edward Kim

cs.CLOct 2, 2026

Query-aware routing for Cross-lingual performance gains in Encoders

Authors: Akshay Jain, Edward Kim

Organizations: ConfidentialMind

Abstract

Multilingual encoders can exhibit reduced retrieval effectiveness when queries and relevant documents differ in language, despite strong same-language performance. We investigate whether Finnish and Swedish cross-lingual retrieval can improve while preserving an encoder's existing same-language performance and document index. We combine a query-only low-rank adapter, trained against frozen document embeddings, with deterministic routing based on query and index languages. Cross-language queries use the adapter, while same-language queries use the original encoder. SampoTron, our fine-tuned low-rank (LoRA) adapter alongwith the Nemotron-3-Embed-1B model, improves average retrieval quality across six English, Finnish, and Swedish directions from 0.241 to 0.291 in normalized discounted cumulative gain (nDCG) at rank ten, a 20.9% relative gain on a sampled financial benchmark. All six cross-lingual directions improve, and routing preserves the original same-language performance, including two full-corpus Finnish evaluations. The approach enables selective cross-language specialization with reusable document embedding vectors.

Figures & tables

Panel of Figure 1 — Figure 1: Training and inference for language-routed query adaptation. Training updates only the query adapter using positive and negative document vectors from the frozen encoder. The dashed arrow denotes adapter updates. At inference, query and index language labels select the original or adapted query path. Both paths share one base model and search the same original document index.

Setting	Value
LoRA rank / scaling / dropout	$64$ / $\alpha=128$ / $0.05$
Target projections	q, k, v, o, gate, up, down
Optimizer / learning rate / weight decay	AdamW / $2\times 10^{-5}$ / $0.01$
Effective / replay batch size	256 / 16
Inverse temperature / served dimensions	20 / 1024
Maximum sequence length / random seed	1024 tokens / 42

Table 1: SampoTron training configuration. Only LoRA parameters are trainable. The document cache contains 90,490 unique texts encoded by the frozen base model.

Figure 2: Final SampoTron evaluation on 194 query identities per direction and a 14,330-passage FIQA corpus sample. Cells show nDCG@10. Rows indicate query language and columns indicate index language. The routed matrix follows the language rule: off-diagonal cells use measured adapter scores, while diagonal cells retain original scores. The right panel gives routed-minus-base differences. All six cross-language directions improve. en: English, fi: Finnish, sv: Swedish.

Configuration	FIQA cross	FIQA same	MIRACL-fi	Mr. TyDi-fi
Original Nemotron-3-Embed-1B	0.2406	0.3303	0.7556	0.7284
Query adapter, ungated	0.2908	0.3361	0.6746	0.6648
SampoTron, routed	0.2908	0.3303	0.7556	0.7284

Table 2: nDCG@10 for the original model, the final adapter applied to all queries, and SampoTron with language routing. FIQA values average six cross-language or three same-language directions on the sampled corpus. The Finnish development evaluations search the full corpora. Routed same-language values follow the original scoring path and are inherited from the base measurements.

Explore similar work

Retrieval Disagreement as a Signal: Diagnosing and Steering Multilingual Dense Embedders
Oct 13, 2026Faeze Ghorbanpour, Shaghayegh Kolli, Jana Diesner +1Multilingual IR Representation Steering
School of Computation, Information and Technology, TU Munich · Munich Center for Machine Learning (MCML)
When Does Mixing Help? Analyzing Query Embedding Interpolation in Multilingual Dense Retrieval
Jun 11, 2026Tongyao Zhu, Chao-Ming Huang, Min-Yen KanMultilingual IR Dense Retrieval
National University of Singapore
The Multilingual Curse at the Retrieval Layer: Evidence from Amharic
May 23, 2026Yosef Worku Alemneh, Kidist Amde Mekonnen, Maarten de RijkeMultilingual IR Low-Resource Language Processing
Independent Researcher · University of Amsterdam

Oct 13, 2026J/K move · Enter open · S save

Retrieval Disagreement as a Signal: Diagnosing and Steering Multilingual Dense Embedders

Faeze Ghorbanpour, Shaghayegh Kolli, Jana Diesner, Alexander Fraser

School of Computation, Information and Technology, TU Munich · Munich Center for Machine Learning (MCML)

Multilingual IR Representation Steering

Open paper AskPDF arXiv

Multilingual dense retrievers place all languages in one shared space so that a question finds relevant documents, regardless of which language they are written in. We show that the same question, asked in different languages, retrieves substantially different documents, and this disagreement is not because of weak language understanding. Each language performs well on its own and ranks candidates consistently with the other languages; divergence appears only when languages share an index. We trace the mechanism: each language keeps its own ranking, but in the shared index, its correct answers are pushed far down, buried under non-relevant documents that are either in the query's own language or in English, depending on the model. We introduce disagreement-guided steering: for each language, we contrast the documents that other languages agree on for a question against those that language retrieves idiosyncratically, and take the difference of their representations as a language-specific direction, added to that language's queries at inference. Against baselines of language-mean removal, subspace removal, and per-language rotation, it most strongly reduces same-language crowding while remaining competitive on recall, trading off against mean removal, which better preserves ranking quality.

Abstract

Explore similar work

Retrieval Disagreement as a Signal: Diagnosing and Steering Multilingual Dense Embedders

When Does Mixing Help? Analyzing Query Embedding Interpolation in Multilingual Dense Retrieval

The Multilingual Curse at the Retrieval Layer: Evidence from Amharic