From db50bfd55fa8ba08b5f7159dc3660d515ed65da5 Mon Sep 17 00:00:00 2001
From: Haoye17 <13693543488@163.com>
Date: Thu, 30 Nov 2023 02:42:08 +0800
Subject: [PATCH] modified: index.html
---
index.html | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/index.html b/index.html
index 92e47ea..b63591e 100644
--- a/index.html
+++ b/index.html
@@ -362,13 +362,13 @@ Abstract
1.4K Fine-Grained and Diverse Human Preference Data: We collect 1.4K pieces of segment-level corrections of human feedback on hallucinations, covering hallucination types including objects (41.2%), positions (20.3%), numbers (16.5%), attributes (10%), actions (5.3%), and others (6.8%).
- High Data Efficiency and Scalability: With just 1.4K annotated data, we achieved 34.8% reduction in model hallucinations. Moreover, the decrease in hallucinations becomes more significant as more data is used.
+ High Data Efficiency and Scalability: With just 1.4K annotated data, we achieved a 34.8% reduction in model hallucinations. Moreover, the decrease in hallucinations becomes more significant as more data used.
Enhanced Performance and Computational Efficiency with DDPO: Our Dense Direct Preference Optimization (DDPO) algorithm can better exploit the fine-grained human feedback, allowing training in under 1 hour on 8 A100 GPUs.
- Outstanding Trustworthiness without Compromising Helpfulness: Our model surpasses existing open-source MLLMs in reducing hallucinations rates, mitigates hallucination from over-generalization, and maintains informativeness.
+ Outstanding Trustworthiness without Compromising Helpfulness: Our model surpasses existing open-source MLLMs in reducing hallucination rates, mitigates hallucination from over-generalization, and maintains informativeness.