Media report · AI-ASSISTED BRIEF
Tencent Yuanbao and King's College London Release GroupAssistBench, First Agent Memory Benchmark Built From Real Group Chats
Report publication · YYSI publication
According to QbitAI, Tencent Yuanbao and King's College London's NLP team released GroupAssistBench, a benchmark built from 50 real group chats and 80,299 messages, with 2,395 tasks testing whether agents recall and correctly apply memory. Across 14 configurations, better recall did not guarantee better service; two lightweight interventions improved open-ended service scores. The paper, code and dataset are public.
Source & context
www.qbitai.com · Read the original source ↗
This AI-assisted brief attributes information to the linked media report. It is not an official announcement or independent YYSI reporting. Consult the original report and its evidence.