SCB: SpeechConversationBench for Evaluating Multi-Turn Reasoning in Speech-to-Speech Models
Paper recorded by Signals 4 on 2026-09-30 in cs.AI. Abstract reproduced from arXiv; link to the original below.
Published 2026-09-30 on arXiv · recorded by Signals 4 on 2026-10-01
Category: cs.AI · 人工智能 · first seen 2026-10-01
Abstract
Speech-to-speech systems must solve tasks whose requirements emerge across conversational turns. We introduce SpeechConversationBench (SCB), a focused evaluation of spoken mathematical reasoning using 103 sharded GSM8K problems. The framework compares the original problem delivered in one turn (full), its concatenated information shards delivered together (concat), and incremental spoken disclosur
Read on arXiv →
Cite this page: SCB: SpeechConversationBench for Evaluating Multi-Turn Reasoning in Speech-to-Speech Models: the #39 most recent of 500 cs.AI papers we have recorded (as of 2026-09-30). Source: Signals 4 (Signals API) — https://data.jiangzhang.ca/signals4/t/papers/scb-speechconversationbench-for-evaluating-multi-turn-reasoning-in-speech-to-spe.html
Free to quote with attribution to “Signals 4 (Signals API)”. Machine-readable:
papers.json
Get 4 AI signals a day by email — free.
Get 4 AI signals a day by email — free