Pith. sign in

REVIEW 3 major objections 3 minor 1 cited by

Temporal User Profiling with LLMs: Balancing Short-Term and Long-Term Preferences for Recommendations

T0 review · 3 major / 3 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read LLM-TUP converts timestamp-split user histories into LLM-generated text, embeds them with BERT, and fuses short- and long-term profiles via attention, claiming to outperform item-embedding averaging on real-world datasets.

desk verdict The abstract describes a plausible LLM-based temporal profiling pipeline, but the supplied full text is an unrelated astro-ph paper, so the central empirical claim is unverifiable from the material at hand. read the letter →

arxiv 2508.08454 v1 pith:F5WNYD5K submitted 2025-08-11 cs.IR cs.AI

Milad Sabouri , Masoud Mansoury , Kun Lin , Bamshad Mobasher This is my paper
classification cs.IRcs.AI
keywords userprofilingtemporalpreferencesshort-termandlong-termlargelanguagemodelsBERTattentionfusioncontent-basedrecommendationinteractiontimestamps
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

The paper proposes LLM-TUP, a user-profiling method for content-based recommenders. It separates each user's interaction history into short-term and long-term segments using timestamps, has a large language model narrate each segment in natural language, embeds those narratives with a pre-trained transformer text encoder (BERT), and fuses the two embeddings with an attention mechanism. The central claim is that this temporally split, text-based profile captures preference dynamics that simple averaging or concatenation of item embeddings misses. On real-world datasets, the paper reports that this pipeline substantially improves recommendation quality over several baselines. If correct, it means a user's shifting tastes can be encoded more faithfully as natural-language descriptions than as aggregate vectors, and the attention weights offer a window into whether recent or long-run preferences dominate.

What carries the argument

The central machinery is the LLM-narration-plus-attention-fusion pipeline: timestamp-aware splitting of history, LLM-generated narrative summaries, BERT embedding of those narratives, and attention-based fusion of the short-term and long-term embeddings into one profile vector. The narrative summaries are the load-bearing innovation: they are supposed to preserve preference semantics that raw item embeddings blur, and the attention fusion is what balances recency against long-run taste.

What would settle it

Take a held-out set of users, build their LLM-TUP profiles, then replace each user's long-term summary with a different user's long-term summary while keeping the short-term. If recommendation quality does not drop dramatically, the long-term text is not carrying the claimed information. Conversely, replace the summaries with a shuffled bag of the same item titles: if quality stays high, the narrative ordering and phrasing are irrelevant and simple embedding pooling would do as well.

Watch

Extended reading notes

Core claim

The paper's central claim is that treating user profiling as a text-generation and fusion problem outperforms treating it as vector aggregation. For each user, the interaction history is split at a timestamp cutoff into a short-term and a long-term portion; an LLM generates a natural-language summary of each portion; a pre-trained BERT encoder converts each summary into an embedding; and an attention mechanism learns to weight the two embeddings into a single user profile. The paper asserts that this profile is semantically richer and temporally aware, yielding substantial improvements over baselines that average or concatenate item embeddings.

Load-bearing premise

The load-bearing premise is that an LLM's natural-language summaries of a user's short-term and long-term interaction histories are faithful and information-preserving enough that their text embeddings outperform the original item embeddings for recommendation.

Editorial extensions

If this is right

  • Content-based recommenders can model shifting tastes without explicit sequence models: just narrate the user's recent and long-run history separately.
  • The attention weights provide an interpretable per-user balance between recency and long-term interest.
  • User profiles become human-readable artifacts that can be inspected, corrected, or turned into explanations.
  • The method lets improvements in LLM summarization and text encoders feed directly into recommendation quality without changing the fusion architecture.

Reading between the lines

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

  • An obvious extension, not specified in the abstract, is to make the short/long cutoff per-user or adaptive, since a fixed timestamp boundary will compress very different interaction cadences into the same two buckets; the effect on the reported gains is testable.
  • If the narrative summaries are the source of the gain, the same text could double as an explanation layer for recommendations, letting a system point to 'your recent interest in X' as justification.
  • A stress test implied by the design is to run the same attention fusion on timestamp-split averaged item embeddings, without any LLM narration; the gap between that variant and LLM-TUP would isolate how much of the improvement comes from language semantics versus the temporal split itself.
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 / 3 minor

Summary. The abstract describes LLM-TUP, a content-based recommendation method that generates natural-language representations of user histories with an LLM, splits them into short-term and long-term segments using interaction timestamps, encodes these with a pre-trained BERT model, and fuses the embeddings with an attention mechanism. The abstract claims that experiments on real-world datasets show substantial improvements over several baselines. The supplied full text, however, is an unrelated astro-ph paper, 'MAGAZ3NE: Far-IR and Radio Insights into the Nature and Properties of Ultramassive Galaxies at z ≳ 3' (arXiv:2508.08460v1), with different authors, title, and content. Therefore, the submitted manuscript contains no method details, no experimental setup, no datasets, no baselines, and no results that could support the abstract's central empirical claim.

Significance. If the claimed results were properly supported, the idea of using LLM-generated, timestamp-split narrative profiles fused with attention would be a plausible and timely contribution to content-based recommendation. The proposal has a clear, falsifiable core hypothesis: LLM/BERT-encoded temporal profiles outperform averaged or concatenated item-embedding profiles. However, the current submission provides no evidence, no reproducibility artifacts, and no formal specification of the LLM prompt, timestamp cutoff, or fusion architecture. The significance of the work therefore cannot be assessed beyond the general plausibility of the approach.

major comments (3)
  1. [Full text] The entire body of the submitted manuscript is an unrelated astronomy paper, 'MAGAZ3NE...' (arXiv:2508.08460v1). It contains no description of LLM-TUP, no datasets, no baselines, no metrics, no protocol, and no results. The abstract's load-bearing sentence, 'Experimental results on real-world datasets demonstrate that LLM-TUP achieves substantial improvements over several baselines,' is therefore unsupported by any evidence in the reviewed material. This is not a local presentation issue; the submission as it stands cannot be evaluated for soundness.
  2. [Abstract] Even at the level of the proposed design, the temporal split is underspecified: there is no statement of whether the short-term/long-term cutoff is global or per-user, how the horizons are chosen, how the LLM prompt is constructed, how the BERT embeddings are computed, or how the attention weights are trained. These choices are load-bearing for reproducibility and for assessing potential data leakage (e.g., whether the profile generation sees target-item information). The missing full text prevents any check of these components.
  3. [Abstract / LLM fidelity] The method assumes that LLM-generated natural-language summaries are faithful and information-preserving representations of the interaction history. No evidence is provided on summary quality, hallucination, or omission of interactions. If the LLM distorts the history, the subsequent temporal split and attention fusion operate on a corrupted substrate, so the claimed improvements cannot be expected to generalize. The submitted material contains no validation of this premise.
minor comments (3)
  1. [Title and metadata] The manuscript title, author list, and arXiv identifier in the full text do not match the submitted abstract. The full text is formatted as an astronomy paper, not as a cs.IR submission.
  2. [Abstract, wording] The phrase 'substantial improvements' is vague. In a proper version of the paper, effect sizes, confidence intervals, and statistical significance should be reported.
  3. [References] No references are provided in the reviewed material, so it is impossible to identify the baselines, datasets, or prior temporal-profiling work against which LLM-TUP should be compared.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found; the abstract's empirical claim is unverified because the supplied full text is an unrelated astro-ph paper, but no step reduces to its own input.

full rationale

The only available material from the cs.IR paper is its abstract. The proposed method (LLM-TUP) generates natural-language summaries of user histories from interaction timestamps, encodes them with BERT, and fuses short-term and long-term embeddings with attention; none of these steps define their output in terms of the claimed experimental result. The single quantitative claim, 'substantial improvements over several baselines,' is an external comparison against baselines, not a quantity constructed from the method's own fitted values. No equations are given, so no self-definitional or fitted-input-called-prediction step can be exhibited. There is no load-bearing self-citation in the reviewed material. The supplied full text is arXiv:2508.08460v1, an astro-ph.GA paper on ultramassive galaxies, not the LLM-TUP paper; this means the experiments, datasets, and protocol supporting the abstract's claim are absent from the reviewed material. That is a serious verification gap, but an evidentiary gap is not circularity. Per the instruction not to manufacture circularity, the appropriate finding is no significant circularity (score 0).

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

Abstract-only audit. The pipeline's load-bearing choices are unstated: the temporal cutoff, the prompt template, and the learned fusion parameters. No new physical or mathematical entities are postulated; the 'user profile' is a composed vector, not an entity with an independent falsifiable handle.

free parameters (3)
  • Temporal split cutoff (short-term vs long-term threshold)
    The method divides histories into short-term and long-term parts using interaction timestamps; the cutoff value, global or per-user, is not specified in the abstract and would be tuned on validation data.
  • LLM prompt template
    The natural-language representation of a user's history depends on a hand-designed prompt; prompt wording is a free design choice that materially affects the resulting BERT embeddings.
  • Attention fusion weights and architecture hyperparameters
    Attention parameters are trained, and hidden dimensions, BERT variant, and training settings are undisclosed design choices that the reported performance depends on.
assumptions (3)
  • domain assumption LLM-generated natural-language summaries of interaction histories are faithful and preserve preference-relevant semantics without hallucinating or omitting key interactions.
    The entire method relies on generated text being a trustworthy substrate for encoding; invoked by the abstract's core step of generating natural-language representations of user histories.
  • domain assumption A single timestamp-based dichotomy (short-term versus long-term) captures the user preference dynamics that matter for recommendation.
    The abstract states the method 'explicitly models short-term and long-term preferences'; the adequacy of a binary temporal split is assumed and no evidence is available at the abstract level.
  • standard math Pre-trained BERT embeddings and the attention mechanism perform as documented in their source literature and preserve the discriminative signal needed for the downstream task.
    Standard machine-learning tooling treated as given from prior literature, consistent with common practice; not derived or verified in this paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Temporal User Profiling with LLMs: Balancing Short-Term and Long-Term Preferences for Recommendations." pith.science (2026). https://pith.science/paper/F5WNYD5K

@misc{pith2026250808454,
  author       = {Pith},
  title        = {Pith review of: Temporal User Profiling with LLMs: Balancing Short-Term and Long-Term Preferences for Recommendations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/F5WNYD5K}},
  note         = {Machine review of arXiv:2508.08454}
}
read the original abstract

Accurately modeling user preferences is crucial for improving the performance of content-based recommender systems. Existing approaches often rely on simplistic user profiling methods, such as averaging or concatenating item embeddings, which fail to capture the nuanced nature of user preference dynamics, particularly the interactions between long-term and short-term preferences. In this work, we propose LLM-driven Temporal User Profiling (LLM-TUP), a novel method for user profiling that explicitly models short-term and long-term preferences by leveraging interaction timestamps and generating natural language representations of user histories using a large language model (LLM). These representations are encoded into high-dimensional embeddings using a pre-trained BERT model, and an attention mechanism is applied to dynamically fuse the short-term and long-term embeddings into a comprehensive user profile. Experimental results on real-world datasets demonstrate that LLM-TUP achieves substantial improvements over several baselines, underscoring the effectiveness of our temporally aware user-profiling approach and the use of semantically rich user profiles, generated by LLMs, for personalized content-based recommendation.

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. SPARK: Search Personalization via Agent-Driven Retrieval and Knowledge-sharing

    cs.AI 2025-12 conditional novelty 4.0 of 10

    SPARK is a proposed framework that coordinates persona-based LLM agents with layered memory and three collaboration protocols to produce personalized search.

Reference graph

Works this paper leans on

74 extracted references · 10 canonical work pages · cited by 1 Pith paper

  1. [2]

    Ashby, M. L. N., Caputi, K. I., Cowley, W., et al. 2018, ApJS, 237, 39, doi: 10.3847/1538-4365/aad4fb

    doi:10.3847/1538-4365/aad4fb 2018

  2. [3]

    J., & Scott, P

    Asplund, M., Grevesse, N., Sauval, A. J., & Scott, P. 2009, ARA&A, 47, 481, doi: 10.1146/annurev.astro.46.060407.145222

    arXiv 2009

  3. [4]

    A., Phillips, M

    Baldwin, J. A., Phillips, M. M., & Terlevich, R. 1981, PASP, 93, 5, doi: 10.1086/130766

    doi:10.1086/130766 1981

  4. [5]

    A., Weibel, A., et al

    Barrufet, L., Oesch, P. A., Weibel, A., et al. 2023, MNRAS, 522, 449, doi: 10.1093/mnras/stad947

    doi:10.1093/mnras/stad947 2023

  5. [6]

    B., Ellis, R

    Belli, S., Newman, A. B., Ellis, R. S., & Konidaris, N. P. 2014, ApJL, 788, L29, doi: 10.1088/2041-8205/788/2/L29

    doi:10.1088/2041-8205/788/2/l29 2014

  6. [7]

    L., et al

    Belli, S., Park, M., Davies, R. L., et al. 2024, Nature, 630, 54, doi: 10.1038/s41586-024-07412-1

    doi:10.1038/s41586-024-07412-1 2024

  7. [8]

    2007, ApJS, 172, 132, doi: 10.1086/520511

    Bertoldi, F., Carilli, C., Aravena, M., et al. 2007, ApJS, 172, 132, doi: 10.1086/520511

    doi:10.1086/520511 2007

  8. [9]

    2019, A&A, 622, A103, doi: 10.1051/0004-6361/201834156

    Boquien, M., Burgarella, D., Roehlly, Y ., et al. 2019, A&A, 622, A103, doi: 10.1051/0004-6361/201834156

    doi:10.1051/0004-6361/201834156 2019

Show all 74 references
  1. [10]

    2003, MNRAS, 344, 1000, doi: 10.1046/j.1365-8711.2003.06897.x

    Bruzual, G., & Charlot, S. 2003, MNRAS, 344, 1000, doi: 10.1046/j.1365-8711.2003.06897.x

    2003

  2. [11]

    2025, ApJ, 981, 25, doi: 10.3847/1538-4357/adaeaf

    Bugiani, L., Belli, S., Park, M., et al. 2025, ApJ, 981, 25, doi: 10.3847/1538-4357/adaeaf

    2025 doi

  3. [12]

    2018, A&A, 620, A16, doi: 10.1051/0004-6361/201732379

    Butler, A., Huynh, M., Delvecchio, I., et al. 2018, A&A, 620, A16, doi: 10.1051/0004-6361/201732379

    2018 doi

  4. [13]

    C., et al

    Calzetti, D., Armus, L., Bohlin, R. C., et al. 2000, ApJ, 533, 682, doi: 10.1086/308692

    2000 doi

  5. [14]

    C., McLure, R

    Carnall, A. C., McLure, R. J., Dunlop, J. S., & Dav´e, R. 2018, MNRAS, 480, 4379, doi: 10.1093/mnras/sty2169

    2018 doi

  6. [15]

    C., McLeod, D

    Carnall, A. C., McLeod, D. J., McLure, R. J., et al. 2023, MNRAS, 520, 3974, doi: 10.1093/mnras/stad369

    2023 doi

  7. [16]

    2003, PASP, 115, 763, doi: 10.1086/376392

    Chabrier, G. 2003, PASP, 115, 763, doi: 10.1086/376392

    2003 doi

  8. [17]

    2017, ApJS, 233, 19, doi: 10.3847/1538-4365/aa97da

    Chang, Y .-Y ., Le Floc’h, E., Juneau, S., et al. 2017, ApJS, 233, 19, doi: 10.3847/1538-4365/aa97da

    2017 doi

  9. [18]

    Charlot, S., & Fall, S. M. 2000, ApJ, 539, 718, doi: 10.1086/309250

    2000 doi

  10. [19]

    2016, ApJ, 819, 62, doi: 10.3847/0004-637X/819/1/62

    Civano, F., Marchesi, S., Comastri, A., et al. 2016, ApJ, 819, 62, doi: 10.3847/0004-637X/819/1/62

    2016 doi

  11. [20]

    L., Barger, A

    Cowie, L. L., Barger, A. J., Hsu, L. Y ., et al. 2017, ApJ, 837, 139, doi: 10.3847/1538-4357/aa60bb da Cunha, E., Charlot, S., & Elbaz, D. 2008, MNRAS, 388, 1595, doi: 10.1111/j.1365-2966.2008.13535.x da Cunha, E., Walter, F., Smail, I. R., et al. 2015, ApJ, 806, 110, doi: 10....

    2017

  12. [21]

    L., Belli, S., Park, M., et al

    Davies, R. L., Belli, S., Park, M., et al. 2024, MNRAS, 528, 4976, doi: 10.1093/mnras/stae327

    2024 doi

  13. [22]

    T., & Li, A

    Draine, B. T., & Li, A. 2007, ApJ, 657, 810, doi: 10.1086/511055

    2007 doi

  14. [23]

    T., Aniano, G., Krause, O., et al

    Draine, B. T., Aniano, G., Krause, O., et al. 2014, ApJ, 780, 172, doi: 10.1088/0004-637X/780/2/172 16 C HANG ET AL

    2014 doi

  15. [24]

    2020a, ApJL, 890, L1, doi: 10.3847/2041-8213/ab5b9f

    Forrest, B., Annunziatella, M., Wilson, G., et al. 2020a, ApJL, 890, L1, doi: 10.3847/2041-8213/ab5b9f

    doi

  16. [25]

    C., Annunziatella, M., et al

    Forrest, B., Marsan, Z. C., Annunziatella, M., et al. 2020b, ApJ, 903, 47, doi: 10.3847/1538-4357/abb819

    doi

  17. [26]

    2022, ApJ, 938, 109, doi: 10.3847/1538-4357/ac8747

    Forrest, B., Wilson, G., Muzzin, A., et al. 2022, ApJ, 938, 109, doi: 10.3847/1538-4357/ac8747

    2022 doi

  18. [27]

    C., Shah, E., et al

    Forrest, B., Lemaux, B. C., Shah, E., et al. 2023, arXiv e-prints, arXiv:2307.15113, doi: 10.48550/arXiv.2307.15113

    2023 doi arXiv

  19. [28]

    C., Shah, E

    Forrest, B., Lemaux, B. C., Shah, E. A., et al. 2024a, ApJ, 971, 169, doi: 10.3847/1538-4357/ad5e78

    doi

  20. [29]

    C., Muzzin, A., et al

    Forrest, B., Cooper, M. C., Muzzin, A., et al. 2024b, ApJ, 977, 51, doi: 10.3847/1538-4357/ad8b1c

    doi

  21. [30]

    E., Dunlop, J

    Geach, J. E., Dunlop, J. S., Halpern, M., et al. 2017, MNRAS, 465, 1789, doi: 10.1093/mnras/stw2721

    2017 doi

  22. [31]

    J., Lutz, D., et al

    Genzel, R., Tacconi, L. J., Lutz, D., et al. 2015, ApJ, 800, 20, doi: 10.1088/0004-637X/800/1/20

    2015 doi

  23. [32]

    A., Schinnerer, E., Leroy, A., et al

    Groves, B. A., Schinnerer, E., Leroy, A., et al. 2015, ApJ, 799, 96, doi: 10.1088/0004-637X/799/1/96

    2015 doi

  24. [33]

    M., Ibar, E., Villanueva, V ., et al

    Hughes, T. M., Ibar, E., Villanueva, V ., et al. 2017, MNRAS, 468, L103, doi: 10.1093/mnrasl/slx033

    2017 doi

  25. [34]

    J., Bonfield, D

    Jarvis, M. J., Bonfield, D. G., Bruce, V . A., et al. 2013, MNRAS, 428, 1281, doi: 10.1093/mnras/sts118

    2013 doi

  26. [35]

    2018, ApJ, 864, 56, doi: 10.3847/1538-4357/aad4af

    Jin, S., Daddi, E., Liu, D., et al. 2018, ApJ, 864, 56, doi: 10.3847/1538-4357/aad4af

    2018 doi

  27. [36]

    D., Leja, J., Conroy, C., & Speagle, J

    Johnson, B. D., Leja, J., Conroy, C., & Speagle, J. S. 2021, ApJS, 254, 22, doi: 10.3847/1538-4365/abef67

    2021 doi

  28. [37]

    M., & Salim, S

    Juneau, S., Dickinson, M., Alexander, D. M., & Salim, S. 2011, ApJ, 736, 104, doi: 10.1088/0004-637X/736/2/104

    2011 doi

  29. [38]

    2013, ApJ, 764, 176, doi: 10.1088/0004-637X/764/2/176

    Juneau, S., Dickinson, M., Bournaud, F., et al. 2013, ApJ, 764, 176, doi: 10.1088/0004-637X/764/2/176

    2013 doi

  30. [39]

    J., Heisler, C

    Kewley, L. J., Heisler, C. A., Dopita, M. A., & Lumsden, S. 2001, ApJS, 132, 37, doi: 10.1086/318944

    2001 doi

  31. [40]

    G., Labb´e, I., et al

    Kriek, M., van Dokkum, P. G., Labb´e, I., et al. 2009, ApJ, 700, 221, doi: 10.1088/0004-637X/700/1/221

    2009 doi

  32. [41]

    R., McKee, C

    Krumholz, M. R., McKee, C. F., & Tumlinson, J. 2009, ApJ, 699, 850, doi: 10.1088/0004-637X/699/1/850 Labb´e, I., van Dokkum, P., Nelson, E., et al. 2023, Nature, 616, 266, doi: 10.1038/s41586-023-05786-2

    2009 doi

  33. [42]

    2013, A&A, 559, A56, doi: 10.1051/0004-6361/201322667 Le Floc’h, E., Aussel, H., Ilbert, O., et al

    Lamastra, A., Menci, N., Fiore, F., et al. 2013, A&A, 559, A56, doi: 10.1051/0004-6361/201322667 Le Floc’h, E., Aussel, H., Ilbert, O., et al. 2009, ApJ, 703, 222, doi: 10.1088/0004-637X/703/1/222

    2013 doi

  34. [43]

    2017, A&A, 607, A110, doi: 10.1051/0004-6361/201730898

    Lisenfeld, U., Alatalo, K., Zucker, C., et al. 2017, A&A, 607, A110, doi: 10.1051/0004-6361/201730898

    2017 doi

  35. [44]

    2018, ApJ, 853, 172, doi: 10.3847/1538-4357/aaa600

    Liu, D., Daddi, E., Dickinson, M., et al. 2018, ApJ, 853, 172, doi: 10.3847/1538-4357/aaa600

    2018 doi

  36. [45]

    2019a, ApJS, 244, 40, doi: 10.3847/1538-4365/ab42da

    Liu, D., Lang, P., Magnelli, B., et al. 2019a, ApJS, 244, 40, doi: 10.3847/1538-4365/ab42da

    doi

  37. [46]

    2019b, ApJ, 887, 235, doi: 10.3847/1538-4357/ab578d

    Liu, D., Schinnerer, E., Groves, B., et al. 2019b, ApJ, 887, 235, doi: 10.3847/1538-4357/ab578d

    doi

  38. [47]

    2011, A&A, 532, A90, doi: 10.1051/0004-6361/201117107

    Lutz, D., Poglitsch, A., Altieri, B., et al. 2011, A&A, 532, A90, doi: 10.1051/0004-6361/201117107

    2011 doi

  39. [48]

    F., Faucher-Gigu`ere, C.-A., et al

    Ma, X., Hopkins, P. F., Faucher-Gigu`ere, C.-A., et al. 2016, MNRAS, 456, 2140, doi: 10.1093/mnras/stv2659

    2016 doi

  40. [49]

    2008, A&A, 488, 463, doi: 10.1051/0004-6361:200809678

    Maiolino, R., Nagao, T., Grazian, A., et al. 2008, A&A, 488, 463, doi: 10.1051/0004-6361:200809678

    2008 doi

  41. [50]

    2010, MNRAS, 408, 2115, doi: 10.1111/j.1365-2966.2010.17291.x

    Mannucci, F., Cresci, G., Maiolino, R., Marconi, A., & Gnerucci, A. 2010, MNRAS, 408, 2115, doi: 10.1111/j.1365-2966.2010.17291.x

    2010

  42. [51]

    2016, ApJ, 827, 150, doi: 10.3847/0004-637X/827/2/150

    Marchesi, S., Civano, F., Salvato, M., et al. 2016, ApJ, 827, 150, doi: 10.3847/0004-637X/827/2/150

    2016 doi

  43. [52]

    C., Marchesini, D., Brammer, G

    Marsan, Z. C., Marchesini, D., Brammer, G. B., et al. 2017, ApJ, 842, 21, doi: 10.3847/1538-4357/aa7206

    2017 doi

  44. [53]

    C., Muzzin, A., Marchesini, D., et al

    Marsan, Z. C., Muzzin, A., Marchesini, D., et al. 2022, ApJ, 924, 25, doi: 10.3847/1538-4357/ac312a

    2022 doi

  45. [54]

    S., Marchesini, D., Brammer, G

    Martis, N. S., Marchesini, D., Brammer, G. B., et al. 2016, ApJL, 827, L25, doi: 10.3847/2041-8205/827/2/L25

    2016 doi

  46. [55]

    2022, ApJ, 926, 37, doi: 10.3847/1538-4357/ac2b9f —

    McConachie, I., Wilson, G., Forrest, B., et al. 2022, ApJ, 926, 37, doi: 10.3847/1538-4357/ac2b9f —. 2025, ApJ, 978, 17, doi: 10.3847/1538-4357/ad8f36

    2022 doi

  47. [56]

    J., Milvang-Jensen, B., Dunlop, J., et al

    McCracken, H. J., Milvang-Jensen, B., Dunlop, J., et al. 2012, A&A, 544, A156, doi: 10.1051/0004-6361/201219507

    2012 doi

  48. [57]

    2013a, ApJ, 777, 18, doi: 10.1088/0004-637X/777/1/18 —

    Muzzin, A., Marchesini, D., Stefanon, M., et al. 2013a, ApJ, 777, 18, doi: 10.1088/0004-637X/777/1/18 —. 2013b, ApJS, 206, 8, doi: 10.1088/0067-0049/206/1/8

    doi

  49. [58]

    B., & Gunn, J

    Oke, J. B., & Gunn, J. E. 1983, ApJ, 266, 713, doi: 10.1086/160817

    1983 doi

  50. [59]

    Pettini, M., & Pagel, B. E. J. 2004, MNRAS, 348, L59, doi: 10.1111/j.1365-2966.2004.07591.x R´emy-Ruyer, A., Madden, S. C., Galliano, F., et al. 2014, A&A, 563, A31, doi: 10.1051/0004-6361/201322803

    2004

  51. [60]

    2020, ApJ, 905, 40, doi: 10.3847/1538-4357/abc7c4

    Saracco, P., Marchesini, D., La Barbera, F., et al. 2020, ApJ, 905, 40, doi: 10.3847/1538-4357/abc7c4

    2020 doi

  52. [61]

    T., Bondi, M., et al

    Schinnerer, E., Sargent, M. T., Bondi, M., et al. 2010, ApJS, 188, 384, doi: 10.1088/0067-0049/188/2/384

    2010 doi

  53. [62]

    2018a, A&A, 611, A22, doi: 10.1051/0004-6361/201731917

    Schreiber, C., Labb´e, I., Glazebrook, K., et al. 2018a, A&A, 611, A22, doi: 10.1051/0004-6361/201731917

    doi

  54. [63]

    2018b, A&A, 618, A85, doi: 10.1051/0004-6361/201833070

    Schreiber, C., Glazebrook, K., Nanayakkara, T., et al. 2018b, A&A, 618, A85, doi: 10.1051/0004-6361/201833070

    doi

  55. [64]

    2014, ApJ, 783, 84, doi: 10.1088/0004-637X/783/2/84

    Scoville, N., Aussel, H., Sheth, K., et al. 2014, ApJ, 783, 84, doi: 10.1088/0004-637X/783/2/84

    2014 doi

  56. [65]

    2016, ApJ, 820, 83, doi: 10.3847/0004-637X/820/2/83

    Scoville, N., Sheth, K., Aussel, H., et al. 2016, ApJ, 820, 83, doi: 10.3847/0004-637X/820/2/83

    2016 doi

  57. [66]

    2017, ApJ, 837, 150, doi: 10.3847/1538-4357/aa61a0 Smolˇci´c, V ., Novak, M., Bondi, M., et al

    Scoville, N., Lee, N., Vanden Bout, P., et al. 2017, ApJ, 837, 150, doi: 10.3847/1538-4357/aa61a0 Smolˇci´c, V ., Novak, M., Bondi, M., et al. 2017, A&A, 602, A1, doi: 10.1051/0004-6361/201628704

    2017 doi

  58. [67]

    Stalevski, M., Fritz, J., Baes, M., Nakos, T., & Popovi´c, L. ˇC. 2012, MNRAS, 420, 2756, doi: 10.1111/j.1365-2966.2011.19775.x

    2012

  59. [68]

    2016, MNRAS, 458, 2288, doi: 10.1093/mnras/stw444 17

    Stalevski, M., Ricci, C., Ueda, Y ., et al. 2016, MNRAS, 458, 2288, doi: 10.1093/mnras/stw444 17

    2016 doi

  60. [69]

    2015, ApJ, 803, 11, doi: 10.1088/0004-637X/803/1/11

    Stefanon, M., Marchesini, D., Muzzin, A., et al. 2015, ApJ, 803, 11, doi: 10.1088/0004-637X/803/1/11

    2015 doi

  61. [70]

    J., Genzel, R., Saintonge, A., et al

    Tacconi, L. J., Genzel, R., Saintonge, A., et al. 2018, ApJ, 853, 179, doi: 10.3847/1538-4357/aaa4b4 Urbano Stawinski, S. M., Cooper, M. C., Forrest, B., et al. 2024, The Open Journal of Astrophysics, 7, 46, doi: 10.33232/001c.120087 Villa-V´elez, J. A., Buat, V ., Theul´e, P....

    2018 doi

  62. [71]

    2019, Nature, 572, 211, doi: 10.1038/s41586-019-1452-4

    Wang, T., Schreiber, C., Elbaz, D., et al. 2019, Nature, 572, 211, doi: 10.1038/s41586-019-1452-4

    2019 doi

  63. [72]

    E., Labb´e, I., van Dokkum, P

    Whitaker, K. E., Labb´e, I., van Dokkum, P. G., et al. 2011, ApJ, 735, 86, doi: 10.1088/0004-637X/735/2/86

    2011 doi

  64. [73]

    2016, ApJ, 827, 74, doi: 10.3847/0004-637X/827/1/74

    Wuyts, E., Wisnioski, E., Fossati, M., et al. 2016, ApJ, 827, 74, doi: 10.3847/0004-637X/827/1/74

    2016 doi

  65. [74]

    2020, MNRAS, 491, 740, doi: 10.1093/mnras/stz3001

    Yang, G., Boquien, M., Buat, V ., et al. 2020, MNRAS, 491, 740, doi: 10.1093/mnras/stz3001

    2020 doi

  66. [75]

    N., et al

    Yang, G., Boquien, M., Brandt, W. N., et al. 2022, ApJ, 927, 192, doi: 10.3847/1538-4357/ac4971

    2022 doi

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.