Research map / Large language models / Language, safety and interpretability
EchoMind: An Interrelated Multi-level Benchmark for Evaluating Empathetic Speech Language Models
Li Zhou, Lutong Yu, You Lyu, Yihang Lin, Zefeng Zhao, Junyi Ao, Yuhao Zhang, Benyou Wang, Haizhou Li
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 changed | Where | Measured | Effect |
|---|---|---|---|
| provide ideal vocal-cue information to the model alongside audio (ablation) | Step-Audio · EchoMind conversation | C_SpeechRel: +1.10 | Improved |
| provide ideal vocal-cue information to the model alongside audio (ablation) | GPT-4o-Audio · EchoMind conversation | VES: +1.08 | Improved |
| provide ideal vocal-cue information to the model alongside audio (ablation) | Qwen2.5-Omni-7B · EchoMind conversation | C_CtxFit: +0.14 | Improved |
| use enhanced prompt telling model to attend to vocal cues (ablation) | 12 SLMs · EchoMind conversation | C_SpeechRel | Mixed |
| add a system prompt (ablation) | GLM-4-voice and VITA-Audio · EchoMind conversation | conversation quality | Hurt |
| use human-recorded speech instead of TTS-synthesized speech (ablation) | top three SLMs · EchoMind-Human | vocal-cue processing performance | Hurt |
| use higher-pitch audio inputs (ablation) | SLMs · EchoMind | task performance | Improved |
| vary speaker gender (male vs female) (ablation) | SLMs · EchoMind | task performance | No clear effect |
| add Think mode to Audio-Flamingo3 (ablation) | Audio-Flamingo3 · EchoMind integrated reasoning | accuracy: 58.80 -> 42.95 | Hurt |
Closest papers by meaning
- EmpathyChat: Structured Cognitive Reasoning in Empathetic Spoken Dialogue (NeurIPS 2026 · similarity 0.92)
- MMSU: A Massive Multi-task Spoken Language Understanding and Reasoning Benchmark (ICLR 2026 · similarity 0.87)
- MME-Emotion: A Holistic Evaluation Benchmark for Emotional Intelligence in Multimodal Large Language Models (ICLR 2026 · similarity 0.87)
- Dynamic-SUPERB Phase-2: A Collaboratively Expanding Benchmark for Measuring the Capabilities of Spoken Language Models with 180 Tasks (ICLR 2025 · similarity 0.86)
- EmotionThinker: Prosody-Aware Reinforcement Learning for Explainable Speech Emotion Reasoning (ICLR 2026 · similarity 0.86)
- VowelPrompt: Hearing Speech Emotions from Text via Vowel-level Prosodic Augmentation (ICLR 2026 · similarity 0.86)
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.
