AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Towards True Speech-to-Speech Models Without Text Guidance

Xingjian Zhao, Zhe Xu, Luozhijie Jin, Yang Wang, Hanfu Chen, Yaozhou Jiang, Ke Chen, Ruixiao Li, Mingshu Chen, Ruiming Wang, Wenbo Zhang, Qinyuan Cheng and 3 more

ICLR 2026 · Large language models › Language, safety and interpretability · approach: Speech and audio language models

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add modality-based layer split (separate final speech/text layers) (ablation)MOSS-Speech pre-training (NF vs NF--NoSplit) · StoryCloze tS.C.accuracy: 77.12 -> 77.66Improved
add modality-based layer split (ablation)MOSS-Speech pre-training · MMLUaccuracy: 60.97 -> 62.11Improved
add frozen pre-training of text backbone (ablation)MOSS-Speech pre-training (FP--Full vs NF) · StoryCloze tS.C.accuracy: 77.66 -> 85.20Improved
add frozen pre-training of text backbone (ablation)MOSS-Speech pre-training · MMLUaccuracy: 62.11 -> 66.50Improved
unfreeze only shared layers or layer-wise instead of all (ablation)FP--Full · StoryCloze / MMLUaccuracy: tS.C. 85.20 vs 83.27; MMLU 66.50 vs 67.26No clear effect
add speech modality with split design and frozen backboneQwen3-8B · MMLUaccuracy: 76.60 -> 67.19Hurt
generate speech directly without text guidanceMOSS-Speech vs GLM-4-Voice · TriviaQA S->Saccuracy: 43.20 -> 28.8Hurt
generate speech directly without text guidanceMOSS-Speech vs GLM-4-Voice · LlamaQA S->Saccuracy: 65.67 -> 63.67No clear effect
make speech encoder fully causal (streaming) fine-tuned from GLM-4-Voice tokenizerspeech tokenizer · Librispeech ASR WER overallWER: 9.17 -> 10.80Hurt
compress chunk size in flow-matching decoder (12.5hz)CosyVoice 2 decoder · Seed-TTS-Eval-ENWER: 4.63 -> 4.14Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy