Pith. sign in

REVIEW 3 major objections 6 minor 44 references

CLIP-CC-Bench: Evaluating Paragraph-Level Video Descriptions in Video-Language Models

T0 review · 3 major / 6 minor · reviewed 2026-08-08 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read CLIP-CC-Bench evaluates paragraph-level video descriptions by aggregating five embedding judges with Borda count and reports a statistically stable ranking of 17 video-language models, with VideoLLaMA3 first at a perfect 80/80 consensus…

desk verdict A useful, honestly scoped benchmark for paragraph-level video description with a transparent protocol — internally stable, but the embedding-based scores are never validated against human judgment, so construct validity is the open gap. read the letter →

arxiv 2608.04302 v1 pith:TY2Y63SK submitted 2026-08-05 cs.CV cs.IRcs.MM

Mukhtiar Ali , Harsh Dubey , Sugam Mishra , Chulwoo Pack This is my paper · ORCID
classification cs.CVcs.IRcs.MM
keywords video-languagemodelslong-formvideodescriptionevaluationbenchmarkembedding-basedBordaaggregationsemanticmatchingparagraph-levelcaptioningCLIP-CC-Bench
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

CLIP-CC-Bench is a new evaluation suite for paragraph-level video description, built from 200 ninety-second movie clips with expert-written references that deliberately omit proper nouns. The paper argues that existing benchmarks—single-sentence captioning, dense captioning, and video QA—do not measure whether a model can produce a coherent, detailed long-form description of a minute-scale video. To fill that gap, the authors score candidate descriptions by cosine similarity in the embedding space of five state-of-the-art text-embedding models, combining paragraph-level and sentence-level matching into a harmonic mean and aggregating per-model ranks by Borda count. Evaluating seventeen video-language models, they report that VideoLLaMA3 achieves a perfect 80/80 Borda score and that the ranking is statistically stable under 1,000 bootstrap resamples. The paper contributes a transparent, reproducible protocol for long-form video description evaluation, together with evidence that current models capture overall narrative structure better than fine-grained detail.

What carries the argument

The load-bearing mechanism is the per-judge harmonic-mean coarse–fine score, $\mathrm{HM\text{-}CF}_j = 2\cdot \mathrm{Coarse}_j\cdot \mathrm{Fine}_j/(\mathrm{Coarse}_j + \mathrm{Fine}_j)$. $\mathrm{Coarse}_j$ is the cosine similarity between the embedding-encoded candidate paragraph and the reference paragraph; $\mathrm{Fine}_j$ is the F1 of best-match cosine similarities between candidate and reference sentences, averaging precision and recall. Five text-embedding models play the role of independent judges, each producing its own rank of the 17 VLMs, and Borda count aggregates those ranks into a consensus ordering that cancels the judges' differing absolute scoring scales. The coarse–fine decomposition is not a mere combination of existing metrics: it separates narrative-level alignment from fine-grained detail coverage and exposes a systematic gap between the two across all evaluated models.

What would settle it

Give a random sample of, say, 50 clips to human raters and have them rank the generated descriptions of two models that the benchmark places adjacent, for example LLaVA-OneVision versus ViLAMP, which tie on Borda 67. If human preference does not match the embedding-based ordering, or correlates no better than chance with the Borda ranking, the benchmark's ranking would be shown to reflect embedding-model agreement rather than description quality. A simpler check is to swap the reference description for one of equal length that describes a different video and verify that the cosine score drops, which would confirm the metric measures content match rather than generic stylistic similarity.

Watch

Extended reading notes

Core claim

The paper's central claim is that paragraph-level video description can be reliably evaluated by treating five long-context text-embedding models as independent judges and aggregating their ranks with Borda count. On the CLIP-CC-Bench dataset, this protocol yields a complete ordering of 17 video-language models, with VideoLLaMA3 ranked first by every judge (Borda 80/80, mean HM-CF 0.67) and LongVA last (Borda 0, mean 0.48). The authors also establish two structural findings: the evaluation is internally stable, since bootstrap resampling preserves the top model in 100% of replicates and keeps every model's 95% rank interval within one position (two for InternVL2), and all models score higher on coarse-grained paragraph similarity than on fine-grained sentence-level F1, indicating that current VLMs describe overall semantics better than specific visual details.

Load-bearing premise

The protocol's validity rests on the assumption that cosine similarity in text-embedding space, aggregated over five embedding models, faithfully reflects how well a generated paragraph describes the video content; the paper does not test this against human ratings and lists sample-based human meta-evaluation as future work.

Editorial extensions

If this is right

  • The benchmark provides a reproducible, standardized protocol: because evaluation scripts, model outputs, and aggregation tools are released, any VLM can be scored on the same 199-clip paired set and compared on equal footing.
  • The consistent coarse–fine gap implies that fine-grained detail coverage is the weakest component of current long-form description, giving a concrete target for training and prompting improvements.
  • Bootstrap analysis shows the ranking is stable at 200 clips, suggesting that reliable system-level comparisons of paragraph description quality do not require thousands of videos.
  • The architecture-family result, where transformer-based multimodal models lead while specialized temporal models rank lower, indicates that temporal modeling alone is not sufficient for narrative description quality.
  • The top model's mean score of 0.67 and fine-grained scores as low as 0.47 quantify the headroom in the task, so future models have a measurable bar to beat.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the embedding-based ranking survives a human meta-evaluation, the same judge-ensemble plus Borda protocol could transfer to other paragraph-level generation tasks, such as image paragraph captioning, story generation, or long-form summarization where reference answers are full paragraphs.
  • Since the protocol does not reward correct temporal or causal ordering, a natural next step is a modified fine-grained score that also matches event sequences; such a metric might reorder models that narrate events in different orders.
  • The deliberate exclusion of proper nouns may disadvantage models that ground descriptions in world knowledge, and the qualitative analysis shows that LongVA's named references were heavily penalized; a companion benchmark retaining names would separate world-knowledge grounding from raw visual understanding.
  • The five embedding judges may share pretraining data, so their effective independence is untested; a diversity check based on agreement on held-out paraphrase pairs would establish whether the ensemble truly behaves as five independent votes.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper introduces CLIP-CC-Bench, a benchmark for paragraph-level video description, built from 200 roughly 90-second movie clips with expert-written, proper-noun-free reference paragraphs. It proposes an automated evaluation protocol in which five LLM-based MTEB embedding models serve as judges, each computing a coarse paragraph-level cosine similarity (Eq. 1) and a fine sentence-level F1 (Eqs. 2–4), combined into a per-judge harmonic mean HM-CF (Eq. 5) and then aggregated over judges by Borda count (Eq. 6). The authors evaluate 17 VLMs, report VideoLLaMA3 as the top model with a perfect Borda score of 80, and support the ranking with inter-judge agreement statistics and bootstrap resampling. They release the data, code, and leaderboard publicly.

Significance. If the proposed metric is a valid proxy for human-judged paragraph description quality, the benchmark would fill a real gap: current video description evaluation is dominated by n-gram metrics, short-context embedding similarity, and unvalidated LLM judges. The paper is transparent and reproducible: the scoring equations are explicit, the full dataset and evaluation scripts are released, the inter-judge agreement and bootstrap analyses (Tables 5–7) are appropriate for internal reliability, and the qualitative appendix documents a pathological failure mode (LongVA on clip 053) that the metric correctly flags. The deliberate removal of proper nouns and cultural references is a thoughtful design choice for measuring visual understanding rather than memorized associations. The central weakness is external validity: the scores are never compared with human judgments or existing metrics, and the five judges are all LLM-based text embedders that may share stylistic biases. Without such validation, the reported Borda ranking may be internally stable but not yet a demonstrated measure of description quality.

major comments (3)
  1. [§4.3–§6] The central construct validity of the benchmark is not established: the HM-CF scores of Eqs. (1)–(5) are never compared with human judgments or with any existing metric. Section 6 explicitly defers "a targeted sample-based human meta-evaluation" to future work. Since the paper's contribution is an evaluation framework for paragraph-level description quality, internal reliability (inter-judge agreement and bootstrap stability) cannot substitute for external validity; the reported Borda ranking could be a stable ranking of the wrong quantity. The authors should add a validation study on a stratified sample of clips and models: human raters rank or rate the outputs per clip, and the resulting human ordering is correlated with HM-CF and Borda scores; comparisons against BLEU/ROUGE/CIDEr/BERTScore and, where feasible, G-VEval/EMScore would also locate the proposed metric relative to existing practice. Without such evidence, the 80/80 Borda claim for VideoLLaMA3 is not yet a claim about description quality.
  2. [§3.3, §6] Each clip is paired with exactly one expert reference description. The coarse and fine scores therefore measure similarity to a single annotator's narration, not coverage of the space of valid descriptions. A model producing a high-quality but stylistically different description will be penalized, and the metric conflates reference coverage with model quality. The paper acknowledges this in Section 6 but does not quantify the effect. For a validation subset (e.g., 30–50 clips), the authors should collect two or more independent references per clip and show that the ranking is stable across choices of reference; alternatively, report the reference-sensitivity of HM-CF (e.g., variance across references for the same (model, clip) pair). This is load-bearing because the leaderboard's meaning depends on the reference being representative rather than idiosyncratic.
  3. [§5.3, Appendix B.2] The five judges are all LLM-based text embedding models and are likely to share pretraining data and stylistic priors; the high pairwise agreement (Spearman 0.96–0.99) may reflect these shared biases rather than convergent validity. The paper itself shows that NV-Embed behaves differently on length mismatches (Appendix B.2: HM-CF drops from 0.566 to 0.460 for mPLUG-Owl3 on clip 053 while other judges are less affected). Borda aggregation averages over these preferences but does not remove a common bias. I ask for a sensitivity analysis: regress HM-CF or its components on surface features (description length, sentence count, type-token ratio, lexical overlap with the reference) and report the residualized ranking; or include at least one judge from a different model family (e.g., a non-LLM sentence encoder or a cross-encoder) to test whether the consensus ranking is robust to judge architecture.
minor comments (6)
  1. [§3.3] The text says "four graduate-student annotators" but the acknowledgments only name two people for annotation; clarify the exact annotator count and role.
  2. [Table 2] Qwen2.5-32B and Qwen2.5-72B have no "Max Frames" entry yet are evaluated on video; specify how these text-only models received video input (e.g., frame sampling plus an adapter), or correct the table.
  3. [Table 3] The judge column header "Nemo-8B" should match the body text "Llama-Embed-Nemotron-8B" (and the reference list) for consistency.
  4. [Figure 1] The caption introduces "HM-CF_j" only after the figure is described; define it as the harmonic mean of coarse and fine scores (Eq. 5) before first use in the caption.
  5. [§3.4, Appendix] References to "table 10" and "section E" appear before the appendix is formally introduced; format as "Table 10" and "Appendix E" at first use for reader navigation.
  6. [§3.2.1] The clips are described as "approximately 90 seconds" in several places; state the exact duration used in processing and evaluation, since the annotation and scoring depend on it.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: HM-CF scores are defined directly from external embedding judges, the Borda ranking is an explicit aggregation definition, and the paper's stability claims are presented as internal consistency rather than external validation.

full rationale

The CLIP-CC-Bench evaluation pipeline is not circular. Each per-model score HM-CF_j is a closed-form function of five externally published, pretrained embedding judges (Eqs. 1-5); no parameter is fitted to the benchmark or to the target ranking, and the Borda aggregation (Eq. 6) is an explicit definition rather than a fitted prediction. The paper makes no load-bearing self-citation: the five embedding models and the MTEB leaderboard are external resources, and the reference list contains no prior work by the present authors. The bootstrap-stability and inter-judge-agreement claims in Section 5.3 are internal-consistency statements computed from the same scoring pipeline, and the paper presents them precisely in those terms rather than as independent truth. The one substantive gap is external validity: Section 6 explicitly concedes that 'a targeted sample-based human meta-evaluation would help confirm the external validity of the automated scores and remains valuable future work.' That is a validity limitation, not circular reasoning, because the benchmark's ranking is not defined by or reduced to human judgments; the central result (VideoLLaMA3's 80/80 Borda score) follows arithmetically from the stated equations and the reported per-judge scores.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The evaluation protocol depends on four unverified domain assumptions: embedding cosine similarity as a proxy for quality, independence of the five embedding judges, fidelity of the GPT-4o grammar-only cleanup, and representativeness of narrative film clips. There are no fitted parameters; the ensemble size and coarse/fine weighting are hand-selected design choices. No new physical or conceptual entities are introduced.

free parameters (2)
  • Number of embedding judges J = 5
    Hand-selected; the five MTEB embeddings are top-ranked models, but the ensemble size is not justified by a formal criterion, and stability is only tested under clip resampling, not under judge-set perturbation.
  • Coarse/fine weighting in HM-CF = equal weights (1:1)
    Equation 5 combines coarse and fine scores with equal weights; no sensitivity analysis over this weighting is reported.
assumptions (4)
  • domain assumption Cosine similarity in LLM embedding space is a valid proxy for paragraph-level description quality.
    Central to Section 4.3; never validated against human judgments, and Section 6 defers human meta-evaluation to future work.
  • domain assumption The five MTEB embedding models are independent judges whose consensus approximates ground truth.
    Section 4.3.3 and Section 5.3; all five are text-embedding models from the same paradigm, so inter-judge agreement measures internal consistency, not external correctness.
  • domain assumption The GPT-4o grammar-only cleanup preserves annotation content.
    Section 3.3 and Appendix E; no verification that cleanup did not add, remove, or reorder content beyond the prompt instruction.
  • domain assumption The Movieclips samples are representative of long-form video content.
    Section 3.1 and Section 6; the benchmark covers only narrative film, with 52% Dialogue/Drama clips, so generalization to other video domains is untested.

how reviews work

0 comments
Cite this review

Pith. "Pith review of CLIP-CC-Bench: Evaluating Paragraph-Level Video Descriptions in Video-Language Models." pith.science (2026). https://pith.science/paper/TY2Y63SK

@misc{pith2026260804302,
  author       = {Pith},
  title        = {Pith review of: CLIP-CC-Bench: Evaluating Paragraph-Level Video Descriptions in Video-Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TY2Y63SK}},
  note         = {Machine review of arXiv:2608.04302}
}
read the original abstract

Benchmarking video-language models has largely focused on short clips and single-sentence metrics, leaving open whether current systems can generate accurate long-form, paragraph-level descriptions. We introduce CLIP-CC-Bench, an evaluation suite for long-form video description built from 5 hours of movie content segmented into 90-second clips, each paired with an expert-written paragraph-style reference. The evaluation suite employs an ensemble of five state-of-the-art LLM-based embedding models to increase reliability and mitigate single-model bias, and applies two complementary methodologies: (i) coarse-grained semantic matching and (ii) fine-grained semantic matching to compare model-generated descriptions against CLIP-CC-Bench references. Using this framework, we evaluate 17 state-of-the-art video-language models and report both their Borda-aggregated rankings and their average scores on CLIP-CC-Bench. We further quantify the protocol's internal reliability through inter-judge agreement and bootstrap ranking stability. We release standardized evaluation scripts, model outputs, and aggregation tools at https://github.com/Multimodal-Intelligence-Lab/CLIP-CC-Bench to support reproducibility. CLIP-CC-Bench provides a practical evaluation framework for long-form video description, filling a gap left by existing short-clip and QA-only benchmarks.

Figures

Figures reproduced from arXiv: 2608.04302 by the authors.

Figure 1
Figure 1. Overview of CLIP-CC-Bench. Top: dataset construction curates [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Distribution of description characteristics in CLIP [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Sampled frames from clip 112. Top: first ten frames. Bottom: last ten frames. Intermediate frames omitted. [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Sampled frames from clip 053. Top: first ten frames. Bottom: last ten frames. Intermediate frames omitted. [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 21 canonical work pages

  1. [1]

    Yauhen Babakhin, Radek Osmulski, Ronay Ak, Gabriel Moreira, Mengyao Xu, Benedikt Schifferer, Bo Liu, and Even Oldridge. 2025. Llama-Embed-Nemotron- 8B: A Universal Text Embedding Model for Multilingual and Cross-Lingual Tasks. arXiv:2511.07025 [cs.CL]

    arXiv 2025

  2. [2]

    Satanjeev Banerjee and Alon Lavie. 2005. METEOR: An Automatic Metric for MT Evaluation with Improved Correlation with Human Judgments. InProceedings of the ACL Workshop on Intrinsic and Extrinsic Evaluation Measures for Machine Translation and/or Summarization

    2005

  3. [3]

    Chen and William B

    David L. Chen and William B. Dolan. 2011. Collecting Highly Parallel Data for Paraphrase Evaluation. InProceedings of the 49th Annual Meeting of the Association for Computational Linguistics

    work page 2011

  4. [4]

    Lin Chen, Xilin Wei, Jinsong Li, Xiaoyi Dong, Pan Zhang, Yuhang Zang, Zehui Chen, Haodong Duan, Bin Lin, Zhenyu Tang, Li Yuan, Yu Qiao, Dahua Lin, Feng Zhao, and Jiaqi Wang. 2024. ShareGPT4Video: Improving Video Understanding and Generation with Better Captions. InNeurIPS

    work page 2024

  5. [5]

    Zhe Chen, Jiannan Wu, Wenhai Wang, Weijie Su, Guo Chen, Sen Xing, Muyan Zhong, Qinglong Zhang, Xizhou Zhu, Lewei Lu, Bin Li, Ping Luo, Tong Lu, Yu Qiao, and Jifeng Dai. 2024. InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Tasks. InCVPR

    work page 2024

  6. [6]

    Kyung-Min Choi, Jihyun Kang, Minseok Kim, and Hyunwoo J. Kim. 2021. Dra- maQA: Character-Centered Video Story Understanding with Hierarchical QA. In AAAI

    work page 2021

  7. [7]

    Chaoyou Fu, Yuhan Dai, Yongdong Luo, Lei Li, Shuhuai Ren, Renrui Zhang, Zihan Wang, Chenyu Zhou, Yunhang Shen, Mengdan Zhang, Peixian Chen, Yanwei Li, Shaohui Lin, Sirui Zhao, Ke Li, Tong Xu, Xiawu Zheng, Enhong Chen, Rongrong Ji, and Xing Sun. 2024. Video-MME: The First-Ever Comprehensive Evaluation Benchmark of Multi-modal LLMs in Video Analysis. arXiv ...

    arXiv 2024

  8. [8]

    Soichiro Fujita, Tsutomu Hirao, Hidetaka Kamigaito, Manabu Okumura, and Masaaki Nagata. 2020. SODA: Story Oriented Dense Video Captioning Evaluation Framework. InECCV

    work page 2020

Show all 44 references
  1. [9]

    Kristen Grauman, Andrew Westbury, Eugene Byrnes, Zachary Chavis, Antonino Furnari, Rohit Girdhar, Jackson Hamburger, Hao Jiang, Miao Liu, Xingyu Liu, Miguel Martin, Tushar Nagarajan, Ilija Radosavovic, Santhosh Kumar Ramakrish- nan, Fiona Ryan, Jayant Sharma, Michael Wray, Men...

  2. [10]

    Xinshuo Hu, Zifei Shan, Xinping Zhao, Zetian Sun, Zhenyu Liu, Dongfang Li, Shaolin Ye, Xinyuan Wei, Qian Chen, Baotian Hu, and Min Zhang. 2025. KaLM-Embedding: Superior Training Data Brings A Stronger Embedding Model. arXiv:2501.01028 [cs.CL]

    2025 arXiv

  3. [11]

    Yunseok Jang, Yale Song, Youngjae Yu, Youngjin Kim, and Gunhee Kim. 2017. TGIF-QA: Toward Spatio-Temporal Reasoning in Visual Question Answering. In CVPR

    2017

  4. [12]

    Ranjay Krishna, Kenji Hata, Freddy Ren, Li Fei-Fei, and Juan Carlos Niebles. 2017. Dense-Captioning Events in Videos. InICCV

    2017

  5. [13]

    Chankyu Lee, Rajarshi Roy, Mengyao Xu, Jonathan Raiman, Mohammad Shoeybi, Bryan Catanzaro, and Wei Ping. 2024. NV-Embed: Improved Techniques for Training LLMs as Generalist Embedding Models. arXiv preprint arXiv:2405.17428

    2024 arXiv

  6. [14]

    Jie Lei, Licheng Yu, Mohit Bansal, and Tamara L. Berg. 2018. TVQA: Localized, Compositional Video Question Answering. InProceedings of the 2018 Conference on Empirical Methods in Natural Language Processing

    2018

  7. [15]

    KunChang Li, Yinan He, Yi Wang, Yizhuo Li, Wenhai Wang, Ping Luo, Yali Wang, Limin Wang, and Yu Qiao. 2023. VideoChat: Chat-Centric Video Understanding. arXiv preprint arXiv:2305.06355

    2023 arXiv

  8. [16]

    Kunchang Li, Yali Wang, Yinan He, Yizhuo Li, Yi Wang, Yi Liu, Zun Wang, Jilan Xu, Guo Chen, Ping Luo, Limin Wang, and Yu Qiao. 2023. MVBench: A Comprehensive Multi-modal Video Understanding Benchmark. arXiv preprint arXiv:2311.17005

    2023 arXiv

  9. [17]

    Zehan Li, Xin Zhang, Yanzhao Zhang, Dingkun Long, Pengjun Xie, and Meishan Zhang. 2023. Towards General Text Embeddings with Multi-stage Contrastive Learning. arXiv:2308.03281 [cs.CL]

    2023 arXiv

  10. [18]

    Chin-Yew Lin. 2004. ROUGE: A Package for Automatic Evaluation of Summaries. InText Summarization Branches Out

    2004

  11. [19]

    Yang Liu, Dan Iter, Yichong Xu, Shuohang Wang, Ruochen Xu, and Chenguang Zhu. 2023. G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment. arXiv preprint arXiv:2303.16634

    2023 arXiv

  12. [20]

    Niklas Muennighoff, Nouamane Tazi, Loïc Magne, and Nils Reimers. 2023. MTEB: Massive Text Embedding Benchmark. arXiv:2210.07316 [cs.CL]

    2023 arXiv

  13. [21]

    Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002. BLEU: a Method for Automatic Evaluation of Machine Translation. InProceedings of the 40th Annual Meeting of the Association for Computational Linguistics

    2002

  14. [22]

    Stefano Pini, Marcella Cornia, Federico Bolelli, Lorenzo Baraldi, Costantino Grana, and Rita Cucchiara. 2019. M-VAD Names: a Dataset for Video Captioning with Naming.Multimedia Tools and Applications78, 24 (2019), 35583–35608

    2019

  15. [23]

    Nils Reimers and Iryna Gurevych. 2019. Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks. InProceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing

    2019

  16. [24]

    Anna Rohrbach, Marcus Rohrbach, and Bernt Schiele. 2014. Coherent Multi- sentence Video Description with Variable Level of Detail. InGerman Conference on Pattern Recognition

    2014

  17. [25]

    Xiaoqian Shen, Yunyang Xiong, Changsheng Zhao, Lemeng Wu, Jun Chen, Chenchen Zhu, Zechun Liu, Fanyi Xiao, Balakrishnan Varadarajan, Florian Schroff, Varun Jampani, Deqing Sun, and Vikas Chandra. 2024. LongVU: Spa- tiotemporal Adaptive Compression for Long Video-Language Unders...

    2024 arXiv

  18. [26]

    Yaya Shi, Xu Yang, Haiyang Xu, Chunfeng Yuan, Bing Li, Weiming Hu, and Zheng- Jun Zha. 2022. EMScore: Evaluating Video Captioning via Coarse-Grained and Fine-Grained Embedding Matching. InCVPR. 17929–17938

    2022

  19. [27]

    Chen Sun, Fabien Baradel, Kevin Murphy, and Cordelia Schmid. 2019. Contrastive Bidirectional Transformer for Temporal Representation Learning. arXiv preprint arXiv:1906.05743

    2019 arXiv

  20. [28]

    Chen Sun, Austin Myers, Carl Vondrick, Kevin Murphy, and Cordelia Schmid

  21. [29]

    Qwen Team. 2025. Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Models. arXiv:2506.05176 [cs.CL]

    2025 arXiv

  22. [30]

    Tony Cheng Tong, Sirui He, Zhiwen Shao, and Dit-Yan Yeung. 2024. G-VEval: A Versatile Metric for Evaluating Image and Video Captions Using GPT-4o. arXiv:2412.13647 [cs.CV]

    2024 arXiv

  23. [31]

    Lawrence Zitnick, and Devi Parikh

    Ramakrishna Vedantam, C. Lawrence Zitnick, and Devi Parikh. 2015. CIDEr: Consensus-based Image Description Evaluation. InCVPR

    2015

  24. [32]

    Paul Vicol, Makarand Tapaswi, Lluis Castrejon, and Sanja Fidler. 2018. MovieGraphs: Towards Understanding Human-Centric Situations from Videos. InCVPR

    2018

  25. [33]

    Xin Wang, Jiawei Wu, Junkang Chen, Lei Li, Yuan-Fang Wang, and William Yang Wang. 2019. VaTeX: A Large-Scale, High-Quality Multilingual Dataset for Video- and-Language Research. InICCV

    2019

  26. [34]

    Haoning Wu, Dongxu Li, Yizhuo Li, Yitian Yuan, Samson Tan, Ping Luo, Wayne Wu, and Ziwei Liu. 2024. LongVideoBench: A Benchmark for Long-context Interleaved Video-Language Understanding. arXiv preprint arXiv:2407.15754

    2024 arXiv

  27. [35]

    Junbin Xiao, Xindi Shang, Angela Yao, and Tat-Seng Chua. 2021. NExT-QA: Next Phase of Question-Answering to Explaining Temporal Actions. InCVPR

    2021

  28. [36]

    Jun Xu, Tao Mei, Ting Yao, and Yong Rui. 2016. MSR-VTT: A Large Video Description Dataset for Bridging Video and Language. InCVPR

    2016

  29. [37]

    Yuan Yao, Tianyu Yu, Ao Zhang, Chongyi Wang, Junbo Cui, Hongji Zhu, Tianchi Cai, Haoyu Li, Weilin Zhao, Zhihui He, Qianyu Chen, Huarong Zhou, Zhensheng Zou, Haoye Zhang, Shengding Hu, Zhi Zheng, Jie Zhou, Jie Cai, Xu Han, Guoyang Zeng, Dahai Li, Zhiyuan Liu, and Maosong Sun. 2...

    2024 arXiv

  30. [38]

    Jiabo Ye, Haiyang Xu, Haowei Liu, Anwen Hu, Ming Yan, Qi Qian, Ji Zhang, Fei Huang, and Jingren Zhou. 2024. mPLUG-Owl3: Towards Long Image-Sequence Understanding in Multi-Modal Large Language Models. arXiv:2408.04840 [cs.CV]

    2024 arXiv

  31. [39]

    Boqiang Zhang, Kehan Li, Zesen Cheng, Zhiqiang Hu, Yuqian Yuan, Guanzheng Chen, Sicong Leng, Yuming Jiang, Hang Zhang, Xin Li, Peng Jin, Wenqi Zhang, Fan Wang, Lidong Bing, and Deli Zhao. 2025. VideoLLaMA 3: Frontier Multi- modal Foundation Models for Image and Video Understan...

    2025 arXiv

  32. [40]

    Weinberger, and Yoav Artzi

    Tianyi Zhang, Varsha Kishore, Felix Wu, Kilian Q. Weinberger, and Yoav Artzi

  33. [41]

    The Three Amigos

    Luowei Zhou, Chenliang Xu, and Jason Corso. 2018. Towards Automatic Learning of Procedures from Web Instructional Videos. InAAAI. EvalMG @SIGIR ’26, July 24, 2026, Melbourne, VIC, Australia Ali et al. A Qualitative Examples To illustrate the behavior of current VLMs on long-fo...

    2018

  34. [2019]

    arXiv preprint arXiv:1904.01766

    VideoBERT: A Joint Model for Video and Language Representation Learn- ing. arXiv preprint arXiv:1904.01766

    1904 arXiv

  35. [2020]

    InInternational Con- ference on Learning Representations

    BERTScore: Evaluating Text Generation with BERT. InInternational Con- ference on Learning Representations

  36. [2022]

    Ego4D: Around the World in 3,000 Hours of Egocentric Video. InCVPR

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.