Pith. sign in

REVIEW 4 major objections 6 minor 47 references

Hansel: Output Length Controlling Framework for Large Language Models

T0 review · 4 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Periodic hidden tokens give LLMs precise word-count control

desk verdict Periodic countdown tokens are a genuinely new finetuning-only length-control trick with strong empirical results, but the mechanism claim is undercut by the absence of an ablation that isolates the periodic tokens from the initial length token. read the letter →

arxiv 2412.14033 v1 pith:QWMW4KAY submitted 2024-12-18 cs.CL cs.LG

Seoha Song , Junhyun Lee , Hyeonmok Ko This is my paper · ORCID
classification cs.CLcs.LG
keywords lengthcontrolspecialtokensfinetuningdecoderlanguagemodelsextrapolationmeanabsoluteerrorsummarizationdialoguegeneration
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 claims that finetuning a decoder LLM on data augmented with hidden special tokens that record the remaining word count every fixed number of words gives the model a far more accurate and general sense of output length than prompting it to hit a word count. On four summarization and dialogue datasets, this method (Hansel) reduces the mean absolute error between generated and target length to a fraction of what prompt-based finetuning achieves, and it keeps errors small even for target lengths never seen during finetuning, such as 130-word summaries or 5-word replies. The practical payoff is that length control becomes a general capability acquired through finetuning alone, without architectural changes, making it applicable to existing large decoder models.

What carries the argument

The central object is the periodic hidden-special-token protocol. During finetuning, each reference output is augmented so a token $|x\rangle\langle y|$ appears at the start, then tokens $|x-1\rangle, |x-2\rangle, \ldots, |0\rangle$ are inserted every $\Delta$ words (excluding the first token), where $x = \lfloor l/\Delta\rfloor$ and $y = l \bmod \Delta$ for reference length $l$. A residual parameter $\delta$ makes some training examples treat the total length as $l-1$ through $l-\delta$, so the model learns not to stop abruptly at $|0\rangle$, and the $N=10$ tokens before $|0\rangle$ are label-masked to avoid teaching incomplete sentences. The protocol is what carries the argument: it gives the model a hard-coded chain-of-thought for counting, and the paper's extrapolation results are its signature.

What would settle it

Decode a Hansel-finetuned model with a target length beyond the training range and record the generated token stream. If the periodic tokens $|x-1\rangle, |x-2\rangle, \ldots, |0\rangle$ do not appear at the expected stride, or are emitted only in the first few steps, then the low mean absolute error cannot be attributed to the periodic counting mechanism, and the extrapolation result would need a different explanation.

Watch

Extended reading notes

Core claim

Hansel's central claim is that periodically injected hidden special tokens during finetuning teach a decoder LLM to track its own progress toward a target length, turning length control into a general counting skill rather than a memorized range of lengths. The tokens encode the remaining word count: the first token $|x\rangle\langle y|$ announces that $x$ full strides of $\Delta$ words plus $y$ words are left, and after every $\Delta$ words the next token $|x-1\rangle, |x-2\rangle, \ldots, |0\rangle$ appears. At inference the model is given the initial token corresponding to the requested length and, according to the claim, reproduces the periodic token stream, which keeps it on track and lets it stop near $|0\rangle$. The paper's experiments report that this yields mean absolute errors of 0.05 to 0.98 words across four datasets and four model architectures, versus 0.11 to 3.62 for prompt-based finetuning, and that the advantage grows sharply for target lengths far from the training distribution.

Load-bearing premise

The method's success depends on the finetuned model spontaneously re-emitting the periodic special tokens during inference at the same stride used in training; the paper infers this from end-to-end length error rather than directly checking the generated token stream.

Editorial extensions

If this is right

  • Length control becomes a general capability: MAE stays below about one word even when targets are far outside the finetuning distribution, whereas prompt-finetuned models diverge rapidly.
  • The method is architecture-agnostic: it works with rotary, ALiBi, learned, and T5-bias positional encodings, so it applies to most pretrained decoder LLMs through finetuning only.
  • Output quality is not sacrificed: ROUGE-L and G-Eval scores for Hansel match or slightly exceed vanilla finetuning across datasets.
  • Hansel eliminates infinite generation in the tested settings, with no infinite loops in 10,000 generations, while vanilla and prompt-finetuned models produced 8.4 and 15.3 per 10,000 respectively.
  • The same protocol can control multiple units at once, such as sentences and words, by using separate token families.

Reading between the lines

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

  • A direct ablation would be to decode with the initial token but strip the periodic tokens as they appear, or vice versa, to isolate how much of the length control comes from the initial announcement versus the periodic checkpoints.
  • Because the initial token acts as a length switch, the same finetuned model could serve both length-controlled and unconstrained generation by simply omitting or including that token, which the paper's blended training suggests.
  • The periodic-token idea could transfer to preference-optimization methods that suffer from overly long outputs, giving direct preference optimization an explicit length channel rather than a length penalty.
  • Testing with character-level or token-level units might reveal different optimal strides, since the paper notes counting difficulty increases from sentences to tokens to words to characters.
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

4 major / 6 minor

Summary. The paper introduces Hansel, a finetuning framework that aims to give decoder-only LLMs accurate control over output length. During finetuning, the reference outputs are augmented with special tokens that encode the remaining word count: an initial token of the form |x>|y> gives the full remaining length, and further tokens |x-1>, |x-2>, ..., |0> are inserted at intervals of Delta words. The same tokens are expected to appear at inference. The authors compare against a prompt-based length-control finetuning baseline (Gretel) and vanilla finetuning across four datasets (CNN/DM, XSum, DailyDialog, MultiWOZ) and four base models (Phi-2, BLOOM, OPT, T5), reporting large reductions in mean absolute error (MAE) of output length, robust extrapolation to target lengths outside the training distribution, no degradation in ROUGE-L or G-Eval scores, and elimination of infinite generation. The paper also reports a small-sample transfer experiment in which 50 Hansel-style instruction examples transfer length control to a new task.

Significance. If the empirical claims hold, Hansel is a simple, architecture-agnostic finetuning recipe for length control that could be useful in applied settings such as summarization and dialogue systems, and the extrapolation result is genuinely interesting. The breadth of the evaluation—four datasets, four positional-encoding families, and multiple target lengths—is a notable strength, and the method does not require architecture changes or pretraining modifications. However, the paper's central explanatory claim, that the periodically emitted countdown tokens are what makes the method work, is not directly tested: the comparison to Gretel changes both the representation of the target length (special token vs. text) and the presence of periodic tokens in the reference. In addition, the reported results are single runs without error bars or significance tests, and the hyperparameters Delta and delta are selected using the same test sets on which the headline results are reported. These issues leave the mechanism and the statistical robustness of the central claim underdetermined.

major comments (4)
  1. [Method: The Hansel dataset / The Hansel framework] The central comparison, Hansel vs. Gretel (Table 1), conflates two variables: the target length is given as a special token (|x>|y>) rather than as text, and the reference outputs contain periodic countdown tokens. The paper never ablates the periodic tokens. To support the claim that 'the model learns a general means of length control' through periodic tokens, the authors should run a control in which the reference contains the initial |x>|y> token but no periodic tokens (or in which the periodic tokens are not used at inference), and compare its MAE to full Hansel. In addition, the paper states that 'the special tokens will also appear while inference' but never reports whether the model actually emits |x-1>, |x-2>, ..., |0> at the expected word offsets, nor how termination relates to |0>. This is directly measurable from generation logs and is the load-bearing link between the training protocol and the extrapolation results. Without such evidence, the MAE gains could be produced by a model that attends to the initial token alone and learns a length-estimation mapping.
  2. [Results: The effect of hyperparameters (Table 4)] The hyperparameters Delta and delta are selected using the same test sets that are later used for the headline results in Tables 1, 2, and 5. The text reads 'we conclude that Delta = 20 and delta = 1 is the best combination of hyperparameters' after comparing test-set MAE in Table 4, and then those values are used for all other reported experiments on the same test sets. This test-set-based selection can inflate the reported improvements. The authors should hold out a validation split for hyperparameter selection, or at minimum report the results for several hyperparameter settings on a validation set and disclose that the test set was used only once.
  3. [Experimental Setup: Metric and all result tables] No experiment is repeated and no confidence intervals or significance tests are provided. In Table 1, for example, the DailyDialog MAE improves from 0.24 (Gretel) to 0.09 (Hansel); with no variance estimate, this difference may be within run-to-run noise. Given the very small absolute MAE values in the dialogue datasets, the authors should report multiple seeds (at least 3) with mean and standard deviation, or a paired significance test across the test set, to establish that the observed improvements are not artifacts of a single run.
  4. [Results: Target length extrapolation (Table 2)] The paper excludes infinite generations from the MAE and quality statistics, noting in 'Results' that such samples are 'counted separately.' However, the number of excluded samples per cell is not reported for Table 2, where the extrapolation comparisons are the strongest claim. The aggregate statement ('8.4 and 15.3 times per 10,000 generations' for vanilla and Gretel, none for Hansel) is insufficient to assess whether the MAE differences at extreme target lengths, e.g., 130 words in DailyDialog (Hansel 2.07 vs. Gretel 18.51), are driven partly by the removal of divergent generations. Please report the per-cell infinite-generation counts for all three methods.
minor comments (6)
  1. [Abstract and Introduction] The term 'hidden special tokens' is never defined precisely. The tokens are visible in the training text and are expected to be emitted at inference; clarify whether 'hidden' means that they are excluded from the natural-language context, or that they are masked in some way during training.
  2. [Method: The Hansel dataset] The notation |x> (equivalently |x>|0>) and the formula for the initial token |floor(l/Delta)>|l mod Delta> is clear in the example, but the general reader would benefit from an explicit statement that the first token encodes the full remaining length and subsequent tokens count down by Delta, with the residual serving as an offset.
  3. [Introduction / Related Work] The claim that 'length control has yet to be achieved even for the larger LLMs' is an overstatement in light of the cited work on length-following in instruction-tuned models (e.g., Yuan et al. 2024) and should be qualified.
  4. [Experimental Setup: Metric] The aggregation of the four G-Eval category scores into a single number in Tables 1 and 5 is not described. Please state whether the average is over the raw category scores (which have different scales, e.g., 1-3 for fluency/engagingness) or over normalized scores.
  5. [Appendix: Ablation Studies] The phrase 'special token's affect on the output distribution' should be 'effect'; the same typo appears elsewhere in the appendix.
  6. [Appendix: Training and Inference Examples] The appendix states that Hansel training uses 20% vanilla examples and 20% Gretel examples in addition to Hansel examples, and that Gretel training also includes 20% vanilla examples. This mixing ratio is an additional free parameter that is not varied or compared with a purely Hansel dataset; please clarify whether the reported Gretel baseline uses exactly the same vanilla mixture, and discuss any effect of this composition choice.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: Hansel's length-control results are evaluated on held-out test outputs and do not reduce to their training inputs.

full rationale

The paper's central claim is that periodically inserted hidden special tokens improve output-length control in decoder LLMs. Every quantitative result (MAE, ROUGE-L, G-Eval) is measured on held-out test sets against human-written references or an external GPT-4 evaluator, so the predictions are not constructed from the fitted quantities. The Hansel augmentation is a deterministic transformation of the training reference length, but the evaluation target—the length of the model's generated output—is independent of that transformation and is not equal to the augmented token counts by definition. The extrapolation experiment targets lengths (5, 20, 50, 80, 130) that are not used to fit any parameter, and the paper reports the model's absolute error against those arbitrary targets. The choices of Delta=20 and delta=1 are standard hyperparameter selection on the evaluated datasets; even if this risks selection bias, it is not a case of a fitted parameter being renamed as a prediction. The strongest skeptical concern is that the periodic tokens' causal role is underdetermined, because the paper never directly verifies that the model emits the countdown tokens at inference or that they, rather than the initial target token, cause the improvement. That concern is about mechanism identification and empirical support, not circularity: the conclusion would not be true by definition even if the mechanism were different. There are no load-bearing self-citations, no imported uniqueness theorems, and no renaming of a known result. Accordingly, the circularity score is 0.

Assumptions & free parameters 5 free parameters · 5 assumptions · 1 invented entities

The paper's central claim rests on the empirical assumption that finetuning teaches the model to emit and follow periodic counting tokens, plus several hand-set hyperparameters (Delta, delta, N, mixing ratios). No code or data release backs the numbers.

free parameters (5)
  • Delta (stride) = 20
    Interval between special tokens; swept over {10, 20, 40} and chosen as the best compromise against token overhead.
  • delta (residual) = 1
    Maximum residual words after |0>; delta=0 gives the lowest MAE but causes abrupt termination, so delta=1 is chosen for output quality.
  • N (label mask length) = 10
    Number of tokens preceding |0> that are masked during training; no sensitivity analysis is reported.
  • residual reassignment fraction = 20%
    Fraction of training examples whose target length is reduced by 1..delta to teach natural finishing after |0>.
  • vanilla/Gretel mixing ratio = 20%
    Fraction of training examples kept in vanilla or Gretel format so the length-control switch can be turned on and off; not ablated.
assumptions (5)
  • domain assumption LLMs are not intrinsic counters
    Motivated by the GPT-3.5 example in Figure 1; if counting were easy, periodic tokens would be unnecessary.
  • domain assumption Finetuning teaches the model to emit periodic special tokens at inference
    The central mechanism; supported only indirectly by end-to-end MAE, never by direct token-emission analysis.
  • domain assumption Word count is a well-defined, learnable unit
    The paper counts words for MAE but never specifies whether splitting is whitespace-based, token-based, or something else.
  • domain assumption ROUGE-L and G-Eval capture generation quality
    Used to support the 'no degradation' claim, with no statistical testing and GPT-4 as the G-Eval judge.
  • ad hoc to paper The chosen Delta and delta transfer across datasets and models
    Hyperparameters were selected on the same datasets used for final evaluation; no held-out validation set is described.
invented entities (1)
  • Periodic hidden special tokens |x>|y> and |x> independent evidence
    purpose: Carry remaining-word-count information to the model at regular intervals during finetuning and inference
    Their behavioral effect is demonstrated through MAE across four models, but no external replication, formal proof, or direct probing of token emission is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hansel: Output Length Controlling Framework for Large Language Models." pith.science (2026). https://pith.science/paper/QWMW4KAY

@misc{pith2026241214033,
  author       = {Pith},
  title        = {Pith review of: Hansel: Output Length Controlling Framework for Large Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QWMW4KAY}},
  note         = {Machine review of arXiv:2412.14033}
}
read the original abstract

Despite the great success of large language models (LLMs), efficiently controlling the length of the output sequence still remains a challenge. In this paper, we propose Hansel, an efficient framework for length control in LLMs without affecting its generation ability. Hansel utilizes periodically outputted hidden special tokens to keep track of the remaining target length of the output sequence. Together with techniques to avoid abrupt termination of the output, this seemingly simple method proved to be efficient and versatile, while not harming the coherency and fluency of the generated text. The framework can be applied to any pre-trained LLMs during the finetuning stage of the model, regardless of its original positional encoding method. We demonstrate this by finetuning four different LLMs with Hansel and show that the mean absolute error of the output sequence decreases significantly in every model and dataset compared to the prompt-based length control finetuning. Moreover, the framework showed a substantially improved ability to extrapolate to target lengths unseen during finetuning, such as long dialog responses or extremely short summaries. This indicates that the model learns the general means of length control, rather than learning to match output lengths to those seen during training.

Figures

Figures reproduced from arXiv: 2412.14033 by the authors.

Figure 1
Figure 1. An example conversation with GPT. We have [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Schematic of the Hansel framework, compared with the vanilla and Gretel scheme. Vanilla is normal fine-tuning and [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The extrapolation of the length control methods with different target lengths. The dashed line (shaded region) indicates [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: The extrapolation of the length control methods with different target lengths. The dashed line (shaded region) indicates [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: The MAE and ROUGE-L for the Phi-2 Hansel [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: This figure explains how the special tokens are changed with different [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 24 canonical work pages

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    L.; Almeida, D.; Altenschmidt, J.; Altman, S.; Anadkat, S.; et al

    Achiam, J.; Adler, S.; Agarwal, S.; Ahmad, L.; Akkaya, I.; Aleman, F. L.; Almeida, D.; Altenschmidt, J.; Altman, S.; Anadkat, S.; et al. 2023. Gpt-4 technical report. arXiv preprint arXiv:2303.08774

    arXiv 2023

  4. [4]

    M.; Firat, O.; Johnson, M.; Lepikhin, D.; Passos, A.; Shakeri, S.; Taropa, E.; Bailey, P.; Chen, Z.; et al

    Anil, R.; Dai, A. M.; Firat, O.; Johnson, M.; Lepikhin, D.; Passos, A.; Shakeri, S.; Taropa, E.; Bailey, P.; Chen, Z.; et al. 2023. Palm 2 technical report. arXiv preprint arXiv:2305.10403

    arXiv 2023

  5. [5]

    X.; Cao, Y.; Foster, G.; Cherry, C.; et al

    Arivazhagan, N.; Bapna, A.; Firat, O.; Lepikhin, D.; Johnson, M.; Krikun, M.; Chen, M. X.; Cao, Y.; Foster, G.; Cherry, C.; et al. 2019. Massively multilingual neural machine translation in the wild: Findings and challenges. arXiv preprint arXiv:1907.05019

    arXiv 2019

  6. [6]

    Dubey, A.; Jauhri, A.; Pandey, A.; Kadian, A.; Al-Dahle, A.; Letman, A.; Mathur, A.; Schelten, A.; Yang, A.; Fan, A.; et al. 2024. The llama 3 herd of models. arXiv preprint arXiv:2407.21783

    arXiv 2024

  7. [7]

    X.; Taori, R.; Zhang, T.; Gulrajani, I.; Ba, J.; Guestrin, C.; Liang, P

    Dubois, Y.; Li, C. X.; Taori, R.; Zhang, T.; Gulrajani, I.; Ba, J.; Guestrin, C.; Liang, P. S.; and Hashimoto, T. B. 2024. Alpacafarm: A simulation framework for methods that learn from human feedback. Advances in Neural Information Processing Systems, 36

    work page 2024

  8. [8]

    Fan, A.; Grangier, D.; and Auli, M. 2018. Controllable Abstractive Summarization. In Proceedings of the 2nd Workshop on Neural Machine Translation and Generation

    work page 2018

Show all 47 references
  1. [9]

    J.; and Durrett, G

    Goyal, T.; Li, J. J.; and Durrett, G. 2022. News summarization and evaluation in the era of gpt-3. arXiv preprint arXiv:2209.12356

    2022 arXiv

  2. [10]

    He, J.; Kryscinski, W.; McCann, B.; Rajani, N.; and Xiong, C. 2022. CTRL sum: Towards Generic Controllable Text Summarization. In Goldberg, Y.; Kozareva, Z.; and Zhang, Y., eds., Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing

    2022

  3. [11]

    M.; Kocisky, T.; Grefenstette, E.; Espeholt, L.; Kay, W.; Suleyman, M.; and Blunsom, P

    Hermann, K. M.; Kocisky, T.; Grefenstette, E.; Espeholt, L.; Kay, W.; Suleyman, M.; and Blunsom, P. 2015. Teaching machines to read and comprehend. Advances in neural information processing systems, 28

    2015

  4. [12]

    Jie, R.; Meng, X.; Shang, L.; Jiang, X.; and Liu, Q. 2023. Prompt-Based Length Controlled Generation with Reinforcement Learning. arXiv preprint arXiv:2308.12030

    2023 arXiv

  5. [13]

    Karpukhin, V.; Oguz, B.; Min, S.; Lewis, P.; Wu, L.; Edunov, S.; Chen, D.; and Yih, W.-t. 2020. Dense Passage Retrieval for Open-Domain Question Answering. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)

    2020

  6. [14]

    Kikuchi, Y.; Neubig, G.; Sasano, R.; Takamura, H.; and Okumura, M. 2016. Controlling Output Length in Neural Encoder-Decoders. In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing

    2016

  7. [15]

    S.; Reid, M.; Matsuo, Y.; and Iwasawa, Y

    Kojima, T.; Gu, S. S.; Reid, M.; Matsuo, Y.; and Iwasawa, Y. 2022. Large language models are zero-shot reasoners. Advances in neural information processing systems, 35

    2022

  8. [16]

    Lewis, M.; Liu, Y.; Goyal, N.; Ghazvininejad, M.; Mohamed, A.; Levy, O.; Stoyanov, V.; and Zettlemoyer, L. 2020. BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension. In Proceedings of the 58th Annual Meeting of the ...

    2020

  9. [17]

    Li, Y.; Bubeck, S.; Eldan, R.; Del Giorno, A.; Gunasekar, S.; and Lee, Y. T. 2023. Textbooks are all you need ii: phi-1.5 technical report. arXiv preprint arXiv:2309.05463

    2023 arXiv

  10. [18]

    Li, Y.; Su, H.; Shen, X.; Li, W.; Cao, Z.; and Niu, S. 2017. DailyDialog: A Manually Labelled Multi-turn Dialogue Dataset. In Proceedings of the Eighth International Joint Conference on Natural Language Processing (Volume 1: Long Papers)

    2017

  11. [19]

    Lin, C.-Y. 2004. ROUGE : A Package for Automatic Evaluation of Summaries. In Text Summarization Branches Out

    2004

  12. [20]

    Liu, P.; Yuan, W.; Fu, J.; Jiang, Z.; Hayashi, H.; and Neubig, G. 2023 a . Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processing. ACM Comput. Surv., 55(9)

    2023

  13. [21]

    Liu, W.; Bai, Y.; Han, C.; Weng, R.; Xu, J.; Cao, X.; Wang, J.; and Cai, X. 2024. Length Desensitization in Directed Preference Optimization. arXiv preprint arXiv:2409.06411

    2024 arXiv

  14. [22]

    Liu, Y.; Iter, D.; Xu, Y.; Wang, S.; Xu, R.; and Zhu, C. 2023 b . G -Eval: NLG Evaluation using Gpt-4 with Better Human Alignment. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing

    2023

  15. [23]

    Liu, Y.; Jia, Q.; and Zhu, K. 2022. Length control in abstractive summarization by pretraining information selection. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)

    2022

  16. [24]

    Loshchilov, I.; and Hutter, F. 2018. Decoupled Weight Decay Regularization. In International Conference on Learning Representations

    2018

  17. [25]

    Lu, J.; Li, J.; An, S.; Zhao, M.; He, Y.; Yin, D.; and Sun, X. 2024. Eliminating Biased Length Reliance of Direct Preference Optimization via Down-Sampled KL Divergence. arXiv preprint arXiv:2406.10957

    2024 arXiv

  18. [26]

    Meng, Y.; Xia, M.; and Chen, D. 2024. Simpo: Simple preference optimization with a reference-free reward. arXiv preprint arXiv:2405.14734

    2024 arXiv

  19. [27]

    u l c ehre, C .; and Xiang, B

    Nallapati, R.; Zhou, B.; dos Santos, C.; Gu\. u l c ehre, C .; and Xiang, B. 2016. Abstractive Text Summarization using Sequence-to-sequence RNN s and Beyond. In Riezler, S.; and Goldberg, Y., eds., Proceedings of the 20th SIGNLL Conference on Computational Natural Language Learning

    2016

  20. [28]

    B.; and Lapata, M

    Narayan, S.; Cohen, S. B.; and Lapata, M. 2018. Don ' t Give Me the Details, Just the Summary! Topic-Aware Convolutional Neural Networks for Extreme Summarization. In Riloff, E.; Chiang, D.; Hockenmaier, J.; and Tsujii, J., eds., Proceedings of the 2018 Conference on Empirical...

    2018

  21. [29]

    Park, R.; Rafailov, R.; Ermon, S.; and Finn, C. 2024. Disentangling length from quality in direct preference optimization. arXiv preprint arXiv:2403.19159

    2024 arXiv

  22. [30]

    A.; and Lewis, M

    Press, O.; Smith, N. A.; and Lewis, M. 2022. Train short, test long: Attention with linear biases enables input length extrapolation. In 10th International Conference on Learning Representations (ICLR 2022)

    2022

  23. [31]

    D.; Ermon, S.; and Finn, C

    Rafailov, R.; Sharma, A.; Mitchell, E.; Manning, C. D.; Ermon, S.; and Finn, C. 2024. Direct preference optimization: Your language model is secretly a reward model. Advances in Neural Information Processing Systems, 36

    2024

  24. [32]

    Raffel, C.; Shazeer, N.; Roberts, A.; Lee, K.; Narang, S.; Matena, M.; Zhou, Y.; Li, W.; and Liu, P. J. 2020. Exploring the limits of transfer learning with a unified text-to-text transformer. The Journal of Machine Learning Research, 21(1): 5485--5551

    2020

  25. [33]

    S.; Xu, C.; Thakker, U.; Sharma, S

    Sanh, V.; Webson, A.; Raffel, C.; Bach, S.; Sutawika, L.; Alyafeai, Z.; Chaffin, A.; Stiegler, A.; Raja, A.; Dey, M.; Bari, M. S.; Xu, C.; Thakker, U.; Sharma, S. S.; Szczechla, E.; Kim, T.; Chhablani, G.; Nayak, N.; Datta, D.; Chang, J.; Jiang, M. T.-J.; Wang, H.; Manica, M.;...

    2022

  26. [34]

    Shaw, P.; Uszkoreit, J.; and Vaswani, A. 2018. Self-Attention with Relative Position Representations. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 2 (Short Papers)

    2018

  27. [35]

    Singhal, P.; Goyal, T.; Xu, J.; and Durrett, G. 2024. A Long Way to Go: Investigating Length Correlations in RLHF . In First Conference on Language Modeling

    2024

  28. [36]

    Su, J.; Ahmed, M.; Lu, Y.; Pan, S.; Bo, W.; and Liu, Y. 2024. Roformer: Enhanced transformer with rotary position embedding. Neurocomputing, 568: 127063

    2024

  29. [37]

    Sutskever, I.; Vinyals, O.; and Le, Q. V. 2014. Sequence to sequence learning with neural networks. Advances in neural information processing systems, 27

    2014

  30. [38]

    Takase, S.; and Okazaki, N. 2019. Positional Encoding to Control Output Sequence Length. In Burstein, J.; Doran, C.; and Solorio, T., eds., Proceedings of the 2019 Conference of the North A merican Chapter of the Association for Computational Linguistics: Human Language Techno...

    2019

  31. [39]

    Touvron, H.; Martin, L.; Stone, K.; Albert, P.; Almahairi, A.; Babaei, Y.; Bashlykov, N.; Batra, S.; Bhargava, P.; Bhosale, S.; et al. 2023. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288

    2023 arXiv

  32. [40]

    N.; Kaiser, .; and Polosukhin, I

    Vaswani, A.; Shazeer, N.; Parmar, N.; Uszkoreit, J.; Jones, L.; Gomez, A. N.; Kaiser, .; and Polosukhin, I. 2017. Attention is all you need. Advances in neural information processing systems, 30

    2017

  33. [41]

    L.; Fan, A.; Akiki, C.; Pavlick, E.; Ili \'c , S.; Hesslow, D.; Castagn \'e , R.; Luccioni, A

    Workshop, B.; Scao, T. L.; Fan, A.; Akiki, C.; Pavlick, E.; Ili \'c , S.; Hesslow, D.; Castagn \'e , R.; Luccioni, A. S.; Yvon, F.; et al. 2022. Bloom: A 176b-parameter open-access multilingual language model. arXiv preprint arXiv:2211.05100

    2022 arXiv

  34. [42]

    Yu, Z.; Wu, Z.; Zheng, H.; XuanYuan, Z.; Fong, J.; and Su, W. 2021. L en A tten: An Effective Length Controlling Unit For Text Summarization. In Zong, C.; Xia, F.; Li, W.; and Navigli, R., eds., Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021

    2021

  35. [43]

    Yuan, W.; Kulikov, I.; Yu, P.; Cho, K.; Sukhbaatar, S.; Weston, J.; and Xu, J. 2024. Following length constraints in instructions. arXiv preprint arXiv:2406.17744

    2024 arXiv

  36. [44]

    Zang, X.; Rastogi, A.; Sunkara, S.; Gupta, R.; Zhang, J.; and Chen, J. 2020. MultiWOZ 2.2: A Dialogue Dataset with Additional Annotation Corrections and State Tracking Baselines. In Proceedings of the 2nd Workshop on Natural Language Processing for Conversational AI

    2020

  37. [45]

    Zhang, J.; Zhao, Y.; Saleh, M.; and Liu, P. 2020. Pegasus: Pre-training with extracted gap-sentences for abstractive summarization. In International Conference on Machine Learning. PMLR

    2020

  38. [46]

    V.; et al

    Zhang, S.; Roller, S.; Goyal, N.; Artetxe, M.; Chen, M.; Chen, S.; Dewan, C.; Diab, M.; Li, X.; Lin, X. V.; et al. 2022 a . Opt: Open pre-trained transformer language models. arXiv preprint arXiv:2205.01068

    2022 arXiv

  39. [47]

    Zhang, Y.; Zhang, X.; Wang, X.; Chen, S.-q.; and Wei, F. 2022 b . Latent prompt tuning for text summarization. arXiv preprint arXiv:2211.01837

    2022 arXiv

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.