Cross-Encoder

Follow this topic Ask about this topic

Momentum

5 papers in the last four weeks, up 25% on the four weeks before. 0.1% of all new papers.

Jul 6Week of Sep 21

Latest papers 37

All topics

Cross-Lingual Alignment for Decoder-Only Models using MoE Routers
Oct 1, 2026Lucas Bandarkar, Clark Peng, Ahmed Haj Ahmed +2Contrastive Alignment Cross-Encoder
University of California, Los Angeles∗ · Haverford College · MILA - Quebec AI Institute & McGill University
PAIQ: Patch-Aligned Semantic Injection via Residual Rotation
Sep 29, 2026Pinze Ren, Yuwei Zhang, Hao Chen +3Vision Encoders Cross-Encoder
Tsinghua University · Nanyang Technological University
Cross-Country Code-Mixing for Generative Recommendation
Sep 24, 2026Yuan Gao, Hao Deng, Haibo Xing +5Cross-Encoder Crossing
Alibaba International Digital Commerce Group Beijing, China · Alibaba International Digital Commerce Group Hangzhou, China
SCOUT: Sim-to-Real Text-Based Person Retrieval by Embedding-Space Prediction over Frozen Video Features
Sep 16, 2026Abdarahmane Traoré, Andy Couturier, Éric HervetFrozen Encoder Cross-Encoder
Embia, Computer Science Department, Faculty of Science, Université de Moncton, Moncton, NB, Canada
Parameter-Efficient Retrievers for Polish and European Languages
Sep 11, 2026Sławomir Dadas, Rafał Poświata, Małgorzata Grębowiec +1Retrieval Layer Retrievers
National Information Processing Institute al. Niepodległości 188b, 00-608 Warsaw, Poland
Improving Cross-Lingual Token Representations by Adding a Pinch of SALT
Sep 9, 2026Guillem RamírezCross-Encoder Motion-Language Alignment
ILCC, University of Edinburgh
E-SENS: Exclusion-Sensitive Penalization for Negative-Constraint Retrieval
Aug 31, 2026Yerang Kim, Jiyoon Myung, Joohyung HanCross-Encoder Quadratic Slack Penalty
Independent Researcher · Lomin
Frozen Memory Is Not Enough: Rethinking External Memory as Extraction
Aug 17, 2026Mingyuan Li, Guangsheng Yu, Xu Wang +1Peak Memory Cross-Encoder
ELLIS Institute Finland · University of Turku · University of Technology Sydney
Listwise Cross-Encoder Fine-Tuning vs. Agentic Instruction Tuning for LLM Rerankers: A Systematic Study in Medical Procedure Reranking
Aug 10, 2026Matan Fainzilber, Shlomit PlavnerCross-Encoder Reranking Reranking
Healthee
Decomposed Entailment for Factuality Checking and Hallucination Detection
Aug 6, 2026Achir Oukelmoun, Nasredine Semmar, Gaël De ChalendarHallucination Detection Large Language Model Reliability
CEA LIST NANO INNOV 2 Bd Thomas Gobert, 91120 Palaiseau, France
Self-Supervised Representation-Guided Generative Dataset Distillation
Aug 4, 2026Mingzhuo Li, Guang Li, Linfeng Ye +4Diffusion-Based Dataset Distillation Dataset Distillation
Hokkaido University, Japan · University of Toronto, Canada · University of Tokyo, Japan
CeQe: Grounding Lexical Retrieval in Semantic Evidence
Aug 1, 2026Adam Kahirov, Umesh Deshpande, Swaminathan SundararamanCross-Encoder Grounding
IBM Research
Beyond Scale and Generation: Understanding Language Model-based Entity Matching
Jul 27, 2026Zeyu Zhang, Xue Li, Iacer Calixto +2Cross-Encoder Language Modeling
University of Amsterdam & AUMC · CWI · University of Amsterdam +1
Ranked by Position: Order Sensitivity as an Exploitable Attack Surface in LLM Listwise Recommenders
Jul 26, 2026Ge Zhang, Jingru Cheng, Huiyuan ChenLarge Language Model Reranking Order Matters
Stanford University · Independent Researcher
Empowering Cross-Domain Sequential Recommendation with Hybrid Tokenization and Serial-Parallel Decoding
Jul 21, 2026Yuxuan Hu, Yuhao Wang, Tianbo Huang +4Sequential Recommendation Sequence Modeling
City University of Hong Kong · Hong Kong, China · ByteDance Inc. +1
Transforming LLMs into Efficient Cross-Encoders via Knowledge Distillation for RAG Reranking
Jul 11, 2026Shreeya Dasa Lakshminath, Shubhan SCross-Encoder Agentic Retrieval-Augmented Generation Systems
Department of Computer Science and Engineering, B.M.S. College of Engineering, Bangalore, India
Modular Foundation Models for Time-Series Perception in Digital Twins
Jul 3, 2026Quang Hung Pham, Ryad Zemouri, Martin Gagnon +1Time Series Foundation Models Wireless Foundation Models
Hydro-Québec · 75, boulevard René-Lévesque Ouest · Montréal, Québec, Canada
KaLM-Reranker-V1: Fast but Not Late Interaction for Compressed Document Reranking
Jun 22, 2026Xinping Zhao, Jiaxin Xu, Ziqi Dai +9Reranking Cross-Encoder
Harbin Institute of Technology (Shenzhen) · Shenzhen Loop Area Institute (SLAI)
CoRe: A Continuously Reward-Finetuned LLM Query Rewriter for Multi-Stage Context-Aware Relevance in Web-Scale Video Search
Jun 12, 2026Yilin Wen, Rong Yang, Xiaojia Chang +11Rewrite Cross-Encoder
TikTok
Pretrained, Frozen, Still Leaking: Auditing Cross-Encoder Attribute Transfer in EEG Foundation Models
Jun 8, 2026Jianwei TaiElectroencephalography Foundation Models Auditory Attention
School of Internet, Anhui University
Interactions Between Crosscoder Features: A Compact Proofs Perspective
Jun 8, 2026Dmitry Manning-Coe, Thomas Read, Anna Soligo +4Unsupervised Dictionary Learning Cross-Encoder
1Anthony J. Leggett Institute for Condensed Matter Theory · 2MATS · 3UK AI Security Institute (AISI) +4
Harnessing Structural Context for Entity Alignment Foundation Models
Jun 4, 2026Xingyu Chen, Yuanning Cui, Zequn Sun +1Hierarchical Semantic Alignment Cross-Encoder
State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China · Nanjing University of Information Science and Technology, Nanjing, China · National Institute of Healthcare Data Science, Nanjing University, Nanjing, China
When Does Complexity Conditioning Help a Frozen Sentence Embedding? A Controlled Study of Per-Sentence and Pair-Level Difficulty Adaptation
Jun 2, 2026Suhwan HwangCross-Encoder Frozen Language Model
Test-Time Training for Zero-Resource Dense Retrieval Reranking
May 31, 2026Shiyan Liu, Yichen LiRetrieval Layer Cross-Encoder Reranking
Huazhong University of Science and Technology · ByteDance
Sense Representations Are Inducible Interfaces
May 27, 2026Jan Christian Blaise Cruz, Alham Fikri AjiSemantic Representations Cross-Encoder
MBZUAI · SEACrowd
ConvMemory: A Lightweight Learned Memory Reranker, a Negative Attribution Result, and a Research-Preview Conflict Editor
May 27, 2026Taiheng PanConversational Memory Cross-Encoder
School of Computing and Information Systems University of Melbourne
The Multilingual Curse at the Retrieval Layer: Evidence from Amharic
May 23, 2026Yosef Worku Alemneh, Kidist Amde Mekonnen, Maarten de RijkeMultilingual Language Models Multilingual
Independent Researcher · University of Amsterdam
BEiTScore: Reference-free Image Captioning Evaluation with an Efficient Cross-Encoder Model
May 20, 2026Gonçalo Gomes, Bruno Martins, Chrysoula ZervaImage Difference Captioning Captions
Instituto Superior Técnico, University of Lisbon · INESC-ID · Instituto de Telecomunicações
Community-Aware Vertex Ordering for Reference-Based Graph Compression: A Cross-Encoder Empirical Study
May 13, 2026Jimmy DubuissonGraph Representations Data Compression Methods
Vantino, Switzerland
Caraman at SemEval-2026 Task 8: Three-Stage Multi-Turn Retrieval with Query Rewriting, Hybrid Search, and Cross-Encoder Reranking
May 12, 2026David-Maximilian Caraman, Gheorghe Cosmin SilaghiRetrievers Cross-Encoder
Babeş-Bolyai University Cluj-Napoca, Romania
Hypencoder Revisited: Reproducibility and Analysis of Non-Linear Scoring for First-Stage Retrieval
Apr 29, 2026Arne Eichholtz, Yongkang Li, Jutte Vijverberg +2Retrieval Layer Cross-Encoder
University of Amsterdam Amsterdam, The Netherlands
UnIte: Uncertainty-based Iterative Document Sampling for Domain Adaptation in Information Retrieval
Apr 28, 2026Jongyoon Kim, Minseong Hwang, Seung-won HwangDomain Adaptation Retrieval Layer
Interdisciplinary Program in Artificial Intelligence, Seoul National University
ResRank: Unifying Retrieval and Listwise Reranking via End-to-End Joint Training with Residual Passage Compression
Apr 24, 2026Xiaojie Ke, Shuai Zhang, Liansheng Sun +6Cross-Encoder Reranking Large Language Model Compression
Qwen Applications Business Group of Alibaba
Cross-Entropy Is Load-Bearing: A Pre-Registered Scope Test of the K-Way Energy Probe on Bidirectional Predictive Coding
Apr 23, 2026Jon-Paul CacioliCross-Entropy Losses Gumbel-Softmax Relaxation
Independent Researcher, Melbourne, Australia
HeadRank: Decoding-Free Passage Reranking via Preference-Aligned Attention Heads
Apr 19, 2026Juyuan Wang, Chenxing Wang, Yuchen Fang +8Cross-Encoder Reranking Relevance
Weixin Group, Tencent, China · 2South China University of Technology, Guangzhou, China
Neither Here Nor There: Cross-Lingual Representation Dynamics of Code-Mixed Text in Multilingual Encoders
Mar 20, 2026Debajyoti Mazumder, Divyansh Pathak, Prashant Kodali +1Multilingual Language Models Multilingual
Indian Institute of Science Education and Research, Bhopal, India · Microsoft Corporation
Cross-Layer Discrete Concept Discovery for Interpreting Language Models
Jun 24, 2025Ankur Garg, Xuemin Yu, Hassan Sajjad +1Vector-Quantized Variational Autoencoder Cross-Encoder

Oct 1, 2026, cs.CLJ/K move · Enter open · S save

Cross-Lingual Alignment for Decoder-Only Models using MoE Routers

Lucas Bandarkar, Clark Peng, Ahmed Haj Ahmed, Aditi Khandelwal+1

University of California, Los Angeles∗ · Haverford College · MILA - Quebec AI Institute & McGill University

Contrastive Alignment Cross-Encoder

Open paper AskPDF arXiv

Cross-lingual contrastive learning has been a core component of multilingual encoder training, but the ability to explicitly align representations is not possible in decoder-only LLMs because of varying multilingual tokenization. However, a growing amount of research suggests that even in LLMs, higher cross-lingual representational alignment leads to improved cross-lingual transfer. In this paper, we propose a novel approach to reimagine cross-lingual contrastive learning given the architectural constraints of modern LLMs. Rather than applying an auxiliary alignment loss on hidden states, we propose using the outputs of the mixture-of-experts (MoE) routers as the target for alignment. Router outputs lend themselves better to pooling over many tokens, enabling more reliable cross-lingual comparisons at the sequence-level. Controlled continual pre-training experiments on four open-source MoEs show that incorporating this routing loss also aligns the underlying hidden representations across languages. Most importantly, this loss improves multilingual performance on our diverse evaluation suite, demonstrating the potential of cross-lingual MoE router alignment.

Momentum

Related topics

Latest papers 37