Pith Number

pith:U6DHRDMH

pith:2021:U6DHRDMH5ZOVSFEGBMUK5UUQWL

not attested not anchored not stored refs resolved

CLIPScore: A Reference-free Evaluation Metric for Image Captioning

Ari Holtzman, Jack Hessel, Maxwell Forbes, Ronan Le Bras, Yejin Choi

CLIP embeddings can score how well a generated caption matches its image without any human reference captions and match human judgments better than metrics that require them.

arxiv:2104.08718 v3 · 2021-04-18 · cs.CV · cs.CL

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{U6DHRDMH5ZOVSFEGBMUK5UUQWL}

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

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

CLIPScore achieves the highest correlation with human judgements, outperforming existing reference-based metrics like CIDEr and SPICE.

C2weakest assumption

That CLIP's representations pretrained on web data provide a robust, general signal of caption quality that transfers across domains without needing task-specific adaptation or references.

C3one line summary

CLIPScore uses a web-pretrained CLIP model to evaluate image captions without references and achieves higher human correlation than CIDEr or SPICE.

References

70 extracted · 70 resolved · 1 Pith anchors

[1] From Images to Sentences through Scene Description Graphs using Commonsense Reasoning and Knowledge 2015 · arXiv:1511.03292

[2] Evaluating clip: towards characterization of broader capabilities and downstream implications.arXiv preprint arXiv:2108.02818 2021

[3] Peter Anderson, Basura Fernando, Mark Johnson, and Stephen Gould. 2016. Spice: Semantic propositional image caption evaluation. In ECCV. Springer 2016

[4] Mikel Artetxe and Holger Schwenk. 2019. Massively multilingual sentence embeddings for zero-shot cross-lingual transfer and beyond. TACL, 7:597--610 2019

[5] Satanjeev Banerjee and Alon Lavie. 2005. METEOR: an automatic metric for mt evaluation with improved correlation with human judgments. In ACL workshop on Evaluation Measures for MT and Summarization 2005

Formal links

2 machine-checked theorem links

Cited by

73 papers in Pith

Selective Timestep Weighting and Advantage-Based Replay for Sample-Efficient Diffusion RLHF

C3-Bench: A Context-Aware Change Captioning Benchmark

Customizing Video Portraits via Identity-ActionDecoupling

Optimizing Visual Generative Models via Distribution-wise Rewards

VOID: Defeating Unauthorized Mimicry in Latent Diffusion Models

Receipt and verification

First computed	2026-07-05T04:08:03.169682Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

a786788d87ee5d5914860b28aed290b2e210266b3d8bd2d4444804119baacdf8

Aliases

arxiv: 2104.08718 · arxiv_version: 2104.08718v3 · doi: 10.48550/arxiv.2104.08718 · pith_short_12: U6DHRDMH5ZOV · pith_short_16: U6DHRDMH5ZOVSFEG · pith_short_8: U6DHRDMH

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/U6DHRDMH5ZOVSFEGBMUK5UUQWL \
  | 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: a786788d87ee5d5914860b28aed290b2e210266b3d8bd2d4444804119baacdf8

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "1f674c0ec38e27587c85a02d6f5bcd24bb575172db7ea4d02f8b4121441f531c",
    "cross_cats_sorted": [
      "cs.CL"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.CV",
    "submitted_at": "2021-04-18T05:00:29Z",
    "title_canon_sha256": "5a05653470e8c550a92d4fe65e0f4ece7b773022769b49c34d79b20ef59fa93f"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2104.08718",
    "kind": "arxiv",
    "version": 3
  }
}