AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

DPRM: A Dual Implicit Process Reward Model in Multi-Hop Question Answering

Xinyi Wang, Yiping Song, Zhiliang Tian, Bo Liu, Tingjin Luo, Minlie Huang

AAAI 2026 · Large language models › Retrieval and knowledge · approach: Explanatory and multiple-choice QA benchmarks · 1 citations (OpenAlex)

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 co-training of CoT-PRM and KG-PRM (ablation)DPRM · CWQF1: 0.5419 -> 0.5988Improved
add co-training of CoT-PRM and KG-PRM (ablation)DPRM · CWQHit@1: 0.6824 -> 0.7301Improved
add co-training of CoT-PRM and KG-PRM (ablation)DPRM · WebQSPF1: 0.7009 -> 0.7228Improved
add iterative reasoning (ablation)DPRM · CWQF1: 0.5824 -> 0.5988Improved
add iterative reasoning (ablation)DPRM · WebQSPF1: 0.7095 -> 0.7228Improved
add iterative reasoning (ablation)DPRM · WebQSPHit@1: 0.8587 -> 0.8715Improved
add both co-training and iteration (ablation)DPRM · CWQHit@1: 0.6692 -> 0.7301Improved
use KG-PRM instead of ImplicitPRM-DPO as reward modelLlama-3.1-8B reasoning · CWQF1: 0.3538 -> 0.4283Improved
use KG-PRM instead of ImplicitPRM-DPO as reward modelLlama-3.1-8B reasoning · CWQHit@1: 0.5650 -> 0.6341Improved
use KG-PRM instead of ImplicitPRM-DPO as reward modelLlama-3.1-8B reasoning · WebQSPHit@1: 0.7266 -> 0.7346Improved
train PRMs on 1,000 pairs instead of full dataDPRM · WebQSPF1: 0.7228 -> 0.6670Hurt
train PRMs on 1,000 pairs instead of full dataDPRM · CWQHit@1: 0.7301 -> 0.7190Hurt

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