Skip to content
OpenTrain AIFor AI Companies

HFEPX · Eval paper review

What Does Anti-Hallucination Training Buy? Saying Less, Cheap Pruning, and the Rest

Lingkai Bu, Qian Gao

Published

Nov 19, 2026

Citations

0

Trust level

Low

Usefulness score

40/100 (Low)

Extraction confidence

45% (Low)

Derived from extracted protocol signals and abstract evidence.

Rater population

Not reported

Signals refreshed

Nov 19, 2026

Should you rely on this paper?

This paper is adjacent to HFEPX scope and is best used for background context, not as a primary protocol reference.

Use this as background context only. Do not make protocol decisions from this page alone.

Best use

Background context only

Use if you need

Background context only.

What to verify

Read the full paper before copying any benchmark, metric, or protocol choices.

Main weakness

The available metadata is too thin to trust this as a primary source.

Human feedback signal
Detected
From extracted signals
Evaluation signal
Weak or implicit
Validate from full paper
Usefulness for eval research
40/100
Adjacent candidate

Treat as adjacent context, not a core eval-method reference.

Abstract

Preference training is the standard recipe for reducing object hallucination in vision-language models, and each new method reports a lower CHAIR score. We ask what that reduction consists of. Every trained system is compared with its base model's caption truncated to the same object recall, which removes the part of the gain that comes from saying less; the remainder is compared with what a deliberately weak, frozen, eleven-feature post-hoc probe fitted on 500 annotated images can reproduce. All criteria were specified in the analysis code and frozen before the outputs they judge existed. For eight public methods on LLaVA-1.5-7B, the probe reproduces at least two thirds of what training removes beyond saying less for 7 of 8 methods on COCO, 8 of 8 on AMBER and 6 of 8 on Object HalBench, with RLAIF-V the consistent exception; a later public method released for both LLaVA-1.5 and Qwen2.5-VL (SENTINEL) meets the same bar on both. On our own DPO checkpoints the accounting holds on Qwen2.5-VL but fails on InternVL2.5 on COCO (2 of 4 training designs). At matched recall, the base model plus the probe beats truncation for 8 of 9 LLaVA-1.5 systems on COCO and 9 of 9 on AMBER, on the two newer backbones, and when realized as actual sentence deletion. The same measure prices training signals: DPO on the model's own samples labelled by the probe or by an object-level CLIP score recovers 0.70 and 0.87 of the beyond-brevity gain of ground-truth-labelled DPO on Qwen2.5-VL (the probe 0.73 on InternVL2.5), both beat length-matched random labels, and CLIP labels buy more than probe labels on COCO. Human judgements show that the sentences the probe deletes are wrong more often than those truncation would delete; a human check of the CHAIR evaluator's precision falls below our pre-set bar, which we report as a limitation.

What we could verify

These are the protocol signals we could actually recover from the available paper metadata. Use them to decide whether this paper is worth deeper reading.

Human Feedback Types

partial

Pairwise Preference, Rlaif Or Synthetic Feedback

Directly usable for protocol triage.

"Preference training is the standard recipe for reducing object hallucination in vision-language models, and each new method reports a lower CHAIR score."

Evaluation Modes

missing

None explicit

Validate eval design from full paper text.

"Preference training is the standard recipe for reducing object hallucination in vision-language models, and each new method reports a lower CHAIR score."

Quality Controls

missing

Not reported

No explicit QC controls found.

"Preference training is the standard recipe for reducing object hallucination in vision-language models, and each new method reports a lower CHAIR score."

Benchmarks / Datasets

missing

Not extracted

No benchmark anchors detected.

"Preference training is the standard recipe for reducing object hallucination in vision-language models, and each new method reports a lower CHAIR score."

Reported Metrics

missing

Not extracted

No metric anchors detected.

"Preference training is the standard recipe for reducing object hallucination in vision-language models, and each new method reports a lower CHAIR score."

Benchmarks and datasets

No benchmark or dataset names were extracted from the available abstract.

Reported metrics

No metric terms were extracted from the available abstract.

Human feedback details
Uses human feedback
Yes
Feedback types
Pairwise Preference, Rlaif Or Synthetic Feedback
Rater population
Not reported
Expertise required
Coding
Evaluation details
Evaluation modes
None
Agentic eval
None
Quality controls
Not reported
Evidence quality
Low
Use this page as
Background context only

Research brief

Metadata summary

Preference training is the standard recipe for reducing object hallucination in vision-language models, and each new method reports a lower CHAIR score.

Based on abstract + metadata only. Check the source paper before making high-confidence protocol decisions.

Key takeaways

  • Preference training is the standard recipe for reducing object hallucination in vision-language models, and each new method reports a lower CHAIR score.
  • Every trained system is compared with its base model's caption truncated to the same object recall, which removes the part of the gain that comes from saying less; the remainder is compared with what a deliberately weak, frozen, eleven-feature post-hoc probe fitted on 500 annotated images can reproduce.
  • All criteria were specified in the analysis code and frozen before the outputs they judge existed.

Researcher actions

  • Compare this paper against nearby papers in the same arXiv category before using it for protocol decisions.
  • Check the full text for explicit evaluation design choices (raters, protocol, and metrics).
  • Use related-paper links to find stronger protocol-specific references.

Caveats

  • Generated from abstract + metadata only; no PDF parsing.
  • Signals below are heuristic and may miss details reported outside the abstract.

Contribution summary

  • Preference training is the standard recipe for reducing object hallucination in vision-language models, and each new method reports a lower CHAIR score.
  • We ask what that reduction consists of.

Why it matters for eval

  • All criteria were specified in the analysis code and frozen before the outputs they judge existed.

Researcher checklist

  • Human feedback protocol is explicit

    Detected: Pairwise Preference, Rlaif Or Synthetic Feedback

  • Evaluation mode is explicit

    No clear evaluation mode extracted.

  • Quality control reporting appears

    No calibration/adjudication/IAA control explicitly detected.

  • Benchmark or dataset anchors are present

    No benchmark/dataset anchor extracted from abstract.

  • Metric reporting is present

    No metric terms extracted.