Pith Number

pith:GDV46QHW

pith:2023:GDV46QHWLB6AGJVNDS7TLFCWY2

not attested not anchored not stored refs resolved

Beyond Hallucinations: Enhancing LVLMs through Hallucination-Aware Direct Preference Optimization

Bin Wang, Conghui He, Jiaqi Wang, Linke Ouyang, Xiaoyi Dong, Zhiyuan Zhao

HA-DPO trains multimodal models to prefer accurate image descriptions over hallucinatory ones by optimizing on paired responses.

arxiv:2311.16839 v2 · 2023-11-28 · cs.CV · cs.CL

Open paper page JSON Open Graph Bundle Merged state What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{GDV46QHWLB6AGJVNDS7TLFCWY2}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

When applied to three mainstream multimodal models, HA-DPO significantly reduced hallucination issues and amplified the models' generalization capabilities. Notably, the MiniGPT-4 model, when enhanced with HA-DPO, demonstrated a substantial improvement: POPE accuracy rose from 51.13% to 86.13% (an absolute improvement of 35%), and the MME score surged from 932.00 to 1326.46 (a relative improvement of 42.32%).

C2weakest assumption

The constructed positive and negative sample pairs are high-quality, style-consistent, and free of new biases that could undermine preference learning or generalization beyond the tested benchmarks.

C3one line summary

HA-DPO reframes hallucination reduction in LVLMs as direct preference optimization over style-consistent positive and negative response pairs, yielding large gains such as 35-point POPE accuracy jumps on MiniGPT-4.

References

79 extracted · 79 resolved · 1 Pith anchors

[1] Brown, Jack Clark, Sam McCandlish, Chris Olah, Benjamin Mann, and Jared Kaplan 2022

[2] The exploration-exploitation dilemma: a multidisci- plinary framework 2014

[3] Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Sub- biah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakan- tan, Pranav Shyam, Girish Sastry, Amanda Askell, Sand- hini Agarwal, Ariel Herbert-V oss, 2020

[4] Christiano, Jan Leike, Tom B 2017

[5] Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, and Steven C. H. Hoi. Instructblip: Towards general- purpose vision-language models wit 2023

Cited by

20 papers in Pith

Deep Pre-Alignment for VLMs

Vision-EKIPL: External Knowledge-Infused Policy Learning for Visual Reasoning

Aligning Modalities in Vision Large Language Models via Preference Fine-tuning

InternLM-XComposer-2.5: A Versatile Large Vision Language Model Supporting Long-Contextual Input and Output

Multimodal Large Language Models with Adaptive Preference Optimization for Sequential Recommendation

Receipt and verification

First computed	2026-05-17T23:38:14.312573Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

30ebcf40f6587c0326ad1cbf359456c6a539c274676e43cc41abba0df6f8f6da

Aliases

arxiv: 2311.16839 · arxiv_version: 2311.16839v2 · doi: 10.48550/arxiv.2311.16839 · pith_short_12: GDV46QHWLB6A · pith_short_16: GDV46QHWLB6AGJVN · pith_short_8: GDV46QHW

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/GDV46QHWLB6AGJVNDS7TLFCWY2 \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: 30ebcf40f6587c0326ad1cbf359456c6a539c274676e43cc41abba0df6f8f6da

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "005fb17ca22012a217f56f74db2d7f620a1da22e2332a44702e9a4daa218764b",
    "cross_cats_sorted": [
      "cs.CL"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.CV",
    "submitted_at": "2023-11-28T14:54:37Z",
    "title_canon_sha256": "cfd50767e98148c085f71f832e663a8ccb28a0b88fa4daa98d3430463eb4a27c"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2311.16839",
    "kind": "arxiv",
    "version": 2
  }
}