Unverified50% confidenceFactExact time
Multi-hop QA is a class of questions that require reasoning across multiple independent evidence fragments to answer, with mainstream benchmarks including HotpotQA, MuSiQue, and 2WikiMultiHopQA.
1
Sources
50%
Confidence
Long-term
Relevance
10/7/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedReAct框架在HotpotQA多跳问答和Fever事实核查基准上同时提升了任务准确率和推理可解释性67% similarUnverifiedThe research methodology selected questions from TriviaQA that models had already answered correctly, appended incorrect answers, and presented them in two ways—a "trusted source version" and a "user expert version"—with the only variable being the identity of the speaker.66% similarUnverifiedTwo widely cited direct-probing benchmarks exhibited saturation, as frontier models generally gave neutral responses to straightforward questions, causing the benchmarks to detect almost no bias.66% similarUnverifiedGlobal queries can answer inductive questions requiring a macro perspective (such as the main themes across an entire document set), a scenario that pure vector retrieval is almost incapable of handling.64% similarUnverifiedMulti-hop queries need to traverse multiple information nodes in a step-by-step manner, and vector retrieval is ill-suited for such scenarios due to its lack of a relational dimension.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/982963API
curl https://kongchang.com/api/v1/knowledge/claims/982963MCP
get_claim(id=982963)