Frozen Language Model

Follow this topic Ask about this topic

Momentum

7 papers in the last four weeks, down 12% on the four weeks before. 0.1% of all new papers.

Jul 6Week of Sep 21

Latest papers 53

All topics

Voices as Handles: Reasoning about Speaker Identity with Frozen Text LLMs
Sep 29, 2026Runqiu Xu, Zhisheng Zheng, David HarwathVoice Agents Speaker
The University of Texas at Austin, Austin, TX, USA
Locating Answer-Correctness Signals in Frozen Large Language Models
Sep 29, 2026Yuansen Liu, Yixuan Tang, Anthony Kum Hoe TungFrozen Language Model Correctness
School of Computing, National University of Singapore
NinaXander: Feasibility and Limits of Composing Frozen Language Models Across Architecture Families via a Shared Latent Space
Sep 29, 2026Takanori Kotama, Shun-ichiro Hayashi, Daichi Mukunoki +2Frozen Language Model
Graduate School of Informatics Nagoya University · Information Technology Center Nagoya University
SQUARE: Structured Quantum Representation Adapters as Compact Quadratic Feature Maps for Frozen Language Models
Sep 29, 2026Emily Jimin Roh, Hyojun Ahn, Hoyeong Lee +4Large Language Model Quantization Frozen Language Model
Korea University · Sookmyung Women’s University · Ulsan National Institute of Science and Technology +2
BOReFT: Manifold Steering of Language Models for Black-box Optimization
Sep 27, 2026Dhruv Agarwal, Rico Angell, Kavitha Srinivas +4Bayesian Optimization Black-Box Optimization
University of Massachusetts Amherst · New York University · IBM Research +1
Minimally Invasive Steering of Language Models
Sep 24, 2026Taha Entesari, Jingyu Zhang, Daniel Khashabi +1Steering Frozen Language Model
Johns Hopkins University
Learning New Words from Unlabeled Test Data in Automatic Speech Recognition
Sep 24, 2026Mengqi Wang, Mark A. Hasegawa-Johnson, Haolong Zheng +1Automatic Speech Recognition Accents
University of Illinois at Urbana-Champaign, USA · Korea Advanced Institute of Science and Technology, South Korea
Learning from Failures: Heterogeneous Graph Memory for Small Language Model Tool-Using Agents
Sep 23, 2026Jiaxing Li, Lei Song, Rui Dong +1Large Language Model Agents Frozen Language Model
School of Computer Science and Engineering, Southeast University, Nanjing, China
GraphSkillAA: Attribution-Guided Skill-Graph Updating with Targeted Validation and Rollback
Sep 17, 2026Ziqiao Shang, Ling-Yue Ge, Lan-Zhe GuoUnsupervised Skill Discovery Frozen Language Model
National Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China · School of Intelligence Science and Technology, Nanjing University, Suzhou, China
Safe Error Correction for Language Models: Frozen-Base Adjustment with Capability Preservation
Sep 14, 2026Gautam KishoreFrozen Language Model Correction
Eulogik
The Router Within: Eliciting Native Skill Routing from a Frozen LLM
Sep 14, 2026Ruishuo Chen, Xun Wang, Yu Chen +2Skills Large Language Model Routing
Institute for Interdisciplinary Information Sciences, Tsinghua University
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
Prompts in the Wild: A Large Analyzed Collection of Transactional Prompts in Code
Aug 13, 2026Victoria Basmov, Yoav Goldberg, Reut TsarfatyPrompt Engineering Frozen Language Model
Bar-Ilan University · Allen Institute for Artificial Intelligence
DIVE: Unlocking Self-Improvement in Frozen Language Models Through Diversity-Driven Skill Evolution
Aug 12, 2026Siheng Xiong, Ali Payani, Oguzhan Gungordu +1Self-Evolving Skill Libraries Frozen Language Model
Georgia Institute of Technology · Cisco Research
LODESTAR: Trustworthy Entropy Is Navigated, Not Merely Measured -- Reinforced Polarizer Keeps a Frozen LLM from Being Confidently Misled by the Wrong Evidence
Aug 12, 2026Po-Jen Ko, Che-Cheng Wu, Hung-Chun Hsu +2Semantic Entropy Large Language Model Responses
1Research Center for Information Technology Innovation, Academia Sinica, Taiwan
ZetaGPT: A Reference Implementation of Positional--Encoding--Free State--Space--Attention Language Models
Aug 10, 2026Róisín LuoPositional Encoding Linear Attention
University of Galway, Ireland
ChronoState: Hidden Elapsed-Time Conditioning for Temporal-State Action Selection in Frozen-Backbone Language Models
Aug 10, 2026Sam Siavoshian, Omar Ramadan, Amir K. Saeed +3Chronos Frozen Language Model
Independent Researcher, California, USA · Johns Hopkins University, Whiting School of Engineering, Baltimore, MD, USA
Deployable Per-Instance Multi-Layer Activation Steering for Large Language Models
Aug 9, 2026Muhammad Faishal Adly Nelwan, Alfan Farizki WicaksonoLinear Activation Steering Model Activations
Faculty of Computer Science Universitas Indonesia
KV-Skill: Forging Expertise in the Model's Native Language
Aug 5, 2026Zhaowei Han, Xiang Zhang, Bing Han +3Qwen3 Frozen Language Model
University of Michigan, Ann Arbor, MI, USA
IRIS: Reusable Identity Representations from Frozen LLMs for Entity Alignment
Jul 28, 2026Xinran Liu, Shengtao Li, Shouqian Shi +2Hierarchical Semantic Alignment Entities
Smarter and Cheaper at Once: Byte-Exact KV-Cache Grafting Turns a Frozen Small Model into a Verified-Knowledge Flywheel
Jul 15, 2026Sietse SchelpeFrozen Language Model Kv-Cache Management
Corbenic AI
Form, Not Content? A Preregistered, Placebo-Controlled Evaluation of Learned Error-Conditioned Self-Repair Through Prompts and Weights in Frozen Small Code Models
Jul 14, 2026Mehmet IscanControlled Study Frozen Language Model
PythaLab, Yıldız Technical University, Istanbul, Turkey
A Quiet Failure in Calibrated Virtual Screening: Marginal Conformal Prediction Under-Covers the Minority Class, and a Class-Conditional Fix Recovers It
Jul 7, 2026Muhammadjon Tursunbadalov, Mustafojon TursunbadalovOnline Conformal Prediction Minority
School of Science and Technology, Champions College Prep, United States
Safe Inference-Time Alignment via Lagrangian Reward Augmentation
Jul 2, 2026Yaswanth Chittepu, Ativ Joshi, Sohini Chintala +1Reward Gradients Frozen Language Model
University of Massachusetts Amherst · Independent Researcher
Black-Box Inference of LLM Architectural Properties with Restrictive API Access
Jul 1, 2026Christopher Ellis, Shreyas Chaudhari, Mei-Yu Wang +3Attacker Large Language Model Large Language Model Tool Use
Carnegie Mellon University Pittsburgh, PA 15213 · Pittsburgh Supercomputing Center
Mural: Transferring LLM knowledge to image generation via Mixture-of-Transformers
Jun 27, 2026Achin Jain, Jie An, Siddharth Chaudhary +1Modern Text-To-Image Models Image Generation
Amazon AGI
What Intermediate Layers Know: Detecting Jailbreaks from Entropy Dynamics
Jun 23, 2026Sofiia Nikolenko, Michele Papucci, Mina Rezaei +1Large Language Model Jailbreaks Token-Level Entropy
LMU Munich, Munich, Germany · relAI – Konrad Zuse School of Excellence in Reliable AI · University of Pisa, Pisa, Italy +2
SoftSkill: Behavioral Compression for Contextual Adaptation
Jun 18, 2026Xijia Tao, Yihua Teng, Xinyu Fu +6Frozen Language Model Natural Language
The University of Hong Kong · Huawei Research · City University of Hong Kong +1
From Sparse Features to Trustworthy Proxies: Certifying SAE-Based Interpretability
Jun 16, 2026Dibyanayan Bandyopadhyay, Asif EkbalInterpretability Model Activations
Department of Computer Science and Engineering, Indian Institute of Technology Patna, Bihar, India.
SV-Detect: AI-generated Text Detection with Steering Vectors
Jun 5, 2026Mikhail Vishnyakov, Tatiana GaintsevaMachine-Generated Text Detection Frozen Language Model
Independent Researcher · Queen Mary University of London
ReverseEOL: Improving Training-free Text Embeddings via Text Reversal in Decoder-only LLMs
Jun 4, 2026Ailiang Lin, Zhuoyun Li, Yusong Wang +3Frozen Language Model Reverse
Institute of Science Tokyo · Tencent
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
Point-in-Time Financial RAG with Frozen LLMs and Market-Feedback Adaptive Retrieval
May 29, 2026Zijie Zhao, Roy E. WelschFinancial Question Answering Frozen Language Model
Massachusetts Institute of Technology · Cambridge, MA, USA
COFT: Counterfactual-Conformal Decoding for Fair Chain-of-Thought Reasoning in Large Language Models
May 28, 2026Arya Fayyazi, Mehdi Kamal, Massoud PedramLarge Language Model Bias Frozen Language Model
Department of Electrical and Computer Engineering, University of Southern California, Los Angeles, California, USA.
SYNAPSE: Neuro-Symbolic Visual Thought-to-Text Decoding via Topological Semantic Denoising
May 27, 2026Akshaj Murhekar, Abhijit MishraElectroencephalography Decoding Neuro-Symbolic Framework
School of Information University of Texas at Austin
Why LLMs Fail at Causal Discovery and How Interventional Agents Escape
May 26, 2026Amartya Roy, Sonali ParbhooLarge Language Models Fail LLM Reasoning Strategies
SIRE, IIT Delhi and Robert Bosch GmbH, India · Imperial College London
LATTE: Forecasting Peer Anchored Preference Trajectories for Personalized LLM Generation
May 26, 2026Jinze Li, Xiaoyan Yang, Shuo Yang +5Large Language Model Personalization Frozen Language Model
The University of Hong Kong · Ant Healthcare, Ant Group
PaP-NF: Probabilistic Long-Term Time Series Forecasting via Prefix-as-Prompt Reprogramming and Normalizing Flows
May 22, 2026Minju Kim, Youngbum HurProbabilistic Forecasting Time Series
Department of Industrial Engineering, Inha University, Incheon, Republic of Korea
PACE: Two-Timescale Self-Evolution for Small Language Model Agents
May 21, 2026Chen Ling, Pei Chen, Albert Guan +4Self-Evolution Large Language Model Agents
Amazon
Adapting the Interface, Not the Model: Runtime Harness Adaptation for Deterministic LLM Agents
May 21, 2026Tianshi Xu, Huifeng Wen, Meng LiAgent Harness Large Language Model Agents
Peking University
EvoStruct: Bridging Evolutionary and Structural Priors for Antibody CDR Design via Protein Language Model Adaptation
May 20, 2026Mansoor Ahmed, Sujin Lee, Umar Khayaz +1Antibody Epitope Prediction
Georgia State University, Atlanta, USA · Georgia Institute of Technology, Atlanta, USA
LLM Benchmark Datasets Should Be Contamination-Resistant
May 19, 2026Ali Al-Lawati, Jason Lucas, Dongwon Lee +1Benchmark Datasets Large Language Model Benchmarks
The Pennsylvania State University, University Park, PA, USA.
The Bicameral Model: Bidirectional Hidden-State Coupling Between Parallel Language Models
May 11, 2026Cedric Flamant, Udaya Ghai, Kanna ShimizuFrozen Language Model
AWS Agentic AI
PYTHALAB-MERA: Validation-Grounded Memory, Retrieval, and Acceptance Control for Frozen-LLM Coding Agents
May 8, 2026Mehmet IscanCode Generation Frozen Language Model
PythaLab, Yildiz Technical University
Stable Agentic Control: Tool-Mediated LLM Architecture for Autonomous Cyber Defense
May 4, 2026Kerri Prinos, Lilianne Brush, Cameron Denton +5Agentic Control Large Language Model Agents
Horizon3.ai San Francisco, CA
Borrowed Geometry: Cross-Distribution Head-Importance Fingerprints of Frozen Pretrained Gemma 4 31B
May 1, 2026Abay BektursunFrozen Language Model Ablation
Independent research
Investigating Linguistic Steering: An Analysis of Adjectival Effects Across Large Language Model Architectures
Apr 28, 2026Lars MalmqvistLinguistics Frozen Language Model
Research and Implementation
Learning Evidence Highlighting for Frozen LLMs
Apr 24, 2026Shaoang Li, Yanhang Shi, Yufei Li +10LLM Reasoning Strategies Frozen Language Model
Stony Brook University · Meta AI
Frozen LLMs as Map-Aware Spatio-Temporal Reasoners for Vehicle Trajectory Prediction
Apr 23, 2026Yanjiao Liu, Jiawei Liu, Xun Gong +1Autonomous Driving Trajectory Forecasting
School of Artificial Intelligence, Jilin University, Changchun 130012, PR China.
LayerTracer: A Joint Task-Particle and Vulnerable-Layer Analysis framework for Arbitrary Large Language Model Architectures
Apr 22, 2026Yuhang Wu, Qinyuan Liu, Qiuyang Zhao +1Layer-Wise Frozen Language Model
China Electronic Technology Nanhu Research Institute, Jiaxing, China
FreezeEmpath: Efficient Training for Empathetic Spoken Chatbots with Frozen LLMs
Apr 20, 2026Yun Hong, Yan Zhou, Yang FengSpeech Language Models Empathy
Key Laboratory of Intelligent Information Processing, Institute of Computing Technology, Chinese Academy of Sciences (ICT/CAS) · State Key Laboratory of AI Safety, Institute of Computing Technology, Chinese Academy of Sciences · University of Chinese Academy of Sciences, Beijing, China
DuaDeep-SeqAffinity: Dual-Branch Deep Learning for Tri-Stream Sequence-Based Antibody--Antigen Affinity Prediction
Dec 26, 2025Aicha Boutorh, Soumia Bouyahiaoui, Manel Kara Laouar +3Antibody Sequence Modeling
National School of Artificial Intelligence (ENSIA) Sidi Abdallah Campus, Algiers, Algeria · Lamine Debaghine University Hospital Center, CHU Bab El Oued, Algiers, Algeria
Dead Weights, Live Signals: Feedforward Graphs of Frozen Language Models
Date pendingMarcus Armstrong, Navid Ayoobi, Arjun MukherjeeFrozen Language Model Feed-Forward
Department of Computer Science University of Houston Houston, TX 77204

Sep 29, 2026, eess.ASJ/K move · Enter open · S save

Voices as Handles: Reasoning about Speaker Identity with Frozen Text LLMs

Runqiu Xu, Zhisheng Zheng, David Harwath

The University of Texas at Austin, Austin, TX, USA

Voice Agents Speaker Frozen Language Model Strong Large Language Model Soft-Token Representations

Open paper AskPDF arXiv

Multi-user voice agents must track who said what across dialogue sessions. Text LLMs are attractive backbones for such agents, but transcripts alone do not expose acoustic speaker identity, leaving the model without a persistent reference for linking information to speakers across sessions. We address this gap by introducing Speaker Handles, soft-token representations that expose acoustic speaker identity to a frozen text LLM for cross-session speaker-dependent reasoning. A three-stage curriculum trains a lightweight projector, with fewer than 0.1% of the backbone's parameters, to map speaker embeddings into these handles. Establishing whether the resulting handles truly support cross-session speaker-dependent reasoning is challenging with existing benchmarks because textual cues can partially reveal fact ownership. We therefore present SpeakerBind, a controlled shared-agent benchmark in which overlapping facts across users require correct cross-session speaker attribution. Speaker Handles achieve 97.40-98.36% accuracy on VoxCeleb1 and 70.40% on SpeakerBind, close to the 71.88% topline. These results show that the proposed Speaker Handles provide an efficient way to integrate acoustic speaker identity into frozen text LLMs for speaker-content reasoning.

Momentum

Related topics

Latest papers 53