Pith. sign in

REVIEW 4 major objections 4 minor 27 references

ICPC: In-context Prompt Compression with Faster Inference

T0 review · 4 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

Pith's one-line read The paper claims that prompt compression can be done with a small pretrained encoder instead of a second large language model, making compression several times faster while matching or slightly exceeding the output quality of LLM-based…

desk verdict Plausible fast compression heuristic, but the paper never measures task performance, so the central claim is unsupported. read the letter →

arxiv 2501.01625 v1 pith:YWMQM5NA submitted 2025-01-03 cs.CL cs.AI

classification cs.CLcs.AI
keywords promptcompressionin-contextlearningmaskedlanguagemodelingtransformerencoderinferencespeedtextredundancyBERTlong-contextefficiency
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 ICPC, a prompt-compression method that uses a small pretrained transformer encoder rather than a large language model to decide which words in a long prompt are redundant. It claims this makes compression several times faster while preserving, and in some comparisons slightly improving, the similarity between the compressed prompt and the original as measured by BLEU, ROUGE, and BERTScore. The motivation is that attention cost grows with prompt length, so shortening prompts before inference should let LLMs handle longer contexts with less memory and latency. The paper aims to establish that encoder-based importance scoring is a viable, lightweight alternative to LLM-based prompt compression.

What carries the argument

The machinery is the information function in Eq. (1), $L(x_i) = \alpha \sum_{n=-k, n \ne 0}^{k} \mathrm{sim}(x_{i+n}, x_i) + \log p(x_i \mid x_{i,k})$, where $x_{i,k}$ is the context window around $x_i$ with $x_i$ itself removed. The first term punishes a word that closely resembles its neighbors, treating it as redundant; the second term gives a low score to words the encoder can predict from context, again marking them as carrying little new information. Words, phrases, or clauses are ranked by $L(x_i)$ and all units at or above the $p$-th percentile of the loss distribution are removed. The percentile threshold makes the amount of deletion adaptive to the input rather than fixed.

What would settle it

Take a fixed set of prompts with downstream answers, delete exactly one token at a time, and measure how much task accuracy drops; if the tokens with the highest $L(x_i)$ from Eq. (1) are the ones whose removal most damages accuracy, the score is measuring predictability rather than task relevance.

Watch

Extended reading notes

Core claim

The paper's central claim is that a pretrained transformer encoder of roughly 110 million parameters can replace a large language model as the prompt compressor. The encoder's masked-language-model probability for each word, combined with the cosine similarity between that word and its local context, is used as an information score; words with the highest scores are deleted. On the paper's own experiments, this yields compressed prompts whose BLEU, ROUGE, and BERTScore values match or slightly exceed those of LLM-based compressors at the same compression ratios, while compression time drops from tens of milliseconds to about 10-17 milliseconds. The paper takes this as evidence that context information from a small encoder suffices for prompt compression, eliminating the memory overhead of running a second large model.

Load-bearing premise

The method assumes that a word is redundant if a pretrained encoder can predict it from its surrounding words and it closely resembles those words, so deleting such words does not remove task-relevant information.

Editorial extensions

If this is right

  • Long prompts can be shortened before the LLM sees them, so attention cost scales with the compressed length rather than the original.
  • The paper reports an increase on BLEU from 42.6 to 43.4 at ratio 0.6, pointing to compression without fidelity loss.
  • Compression time drops from tens of milliseconds to about 10-17 milliseconds, roughly a 3-4x speedup over LLM-based compressors.
  • The same formula works across six different encoder families, so the method is not tied to one model's pretraining scheme.

Reading between the lines

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

  • A natural testable extension is to measure end-to-end latency and downstream task accuracy after the compressed prompt is fed to the target LLM; the paper reports compression time in isolation.
  • Because the deletion threshold is a global percentile over a chunk, an adaptive per-sentence threshold might better protect rare but essential tokens.
  • The local window $k$ in Eq. (1) cannot see cross-sentence dependencies; coupling the score with a coarse sentence-level salience model could extend the method to documents whose key information spans many sentences.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The paper introduces ICPC, a prompt-compression method that uses a pretrained transformer encoder to score each lexical unit by a weighted combination of its masked-language-model probability and cosine similarity to surrounding words, then removes units whose loss exceeds a percentile threshold. The authors evaluate compression fidelity with BLEU, ROUGE, METEOR, and BERTScore on Wikipedia, arXiv, and Reddit passages, compare several encoder backbones, and report compression-time measurements. They conclude that ICPC improves performance and speed on NLP tasks.

Significance. The speed improvement (Table 2: 10.3 ms vs 45.2 ms at ratio 0.8 for LLMLingua) is a plausible practical benefit if it holds across hardware. However, the paper's central claim of better task performance is not tested: all reported quality metrics measure similarity to the original text, not success on downstream tasks. The scoring function behind the method is also asserted rather than validated. As a result, the contribution currently reduces to a fast lossy text filter whose effect on actual prompt-based tasks is unknown. There is no code release, and the method is not fully specified (e.g., the weight alpha is never reported), which further limits reproducibility.

major comments (4)
  1. [Abstract, §5, §4.2, Tables 1 and 3] The abstract and conclusion state that ICPC achieves 'better performance' on 'different types of NLP tasks,' but no downstream task is evaluated. Tables 1 and 3 report BLEU, ROUGE, METEOR, and BERTScore between the compressed and original texts; these are compression-fidelity metrics. The corpora in §4.1.1 are unlabeled long-text collections, and §4.1.1 explicitly justifies dropping SQuAD and PIQA because they are 'not appropriate'—yet without such task benchmarks, the claim that compressed prompts improve task performance does not follow. Please add evaluations on downstream tasks such as question answering, classification, or summarization, or revise the claims to compression fidelity only.
  2. [§3.2, Eq. (1)] The loss L(x_i) is a linear combination of cosine similarity and log MLM probability, but no derivation or independent validation is provided. The method's behavior rests entirely on this ranking; if it is wrong, the compressor deletes tokens needed by the downstream task. Add ablations against each component, a probability-only baseline, and ideally a check that high-L removal preserves task accuracy.
  3. [§4.1, §3.3] Algorithm hyperparameters are not reported: the scalar alpha in Eq. (1), the context window k, the lexical-unit granularity, and the percentile p are never given numerical values. The text says parameters are tuned (§4.1) but does not say on which data or to which values; if tuned on the same corpora used for the reported comparison, the small improvements over LLMLingua could be due to overfitting rather than method quality.
  4. [§4.2, Table 1] The observed differences over LLMLingua are small (e.g., BLEU 42.7 vs 42.1 at ratio 0.8; METEOR 45.3 vs 45.2), and no error bars, confidence intervals, or significance tests are given despite a statement about 'multiple runs' in §4.1. Moreover, the sentence 'ICPC saw the BLEU score increase with ratio 0.6 on metric BLEU from 42.6 to 43.4' misreads the table: 42.6 and 43.4 are METEOR scores for Selective Context and ICPC, respectively, while the BLEU scores at ratio 0.6 are 37.8 (LLMLingua) and 38.0 (ICPC).
minor comments (4)
  1. [Table 2] The column header 'Training time (ms)' conflicts with the caption and text, which describe compression time; please correct this inconsistency.
  2. [§3.1] The term 'participle' (as in 'participle-based filtering') is nonstandard; if segmentation is meant, use 'segmenting' or 'parsing.'
  3. [Figure 1] Figure 1 is referenced in §4.4 but appears missing or incomplete in the manuscript; if it is meant to be included, ensure it is rendered.
  4. [Appendix A.1, Table 6] Approximate inference times for BERT vs GPT-3 are given without specifying hardware; since the speed claim is a core advantage, provide a direct measured comparison on the same hardware.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity; similarity-based evaluation aligns with the compression objective but no derivation step reduces to its inputs.

full rationale

The derivation chain is self-contained and non-circular. Equation (1) defines a deletion score from a pretrained encoder's MLM probability log p(xi | xi,k) plus the average cosine similarity to its context, and the threshold rule in Eqs. (3)-(4) selects tokens below a percentile of that score. Nothing in this chain is fitted to the reported BLEU/ROUGE/BERTScore values, and no fitted parameter is renamed as a prediction. The comparisons against Random Deletion, Selective Context, and LLMLingua are external baselines, and the cited references, including the authors' own survey [2], are not load-bearing for the method's construction. The only mild concern is that the evaluation metrics measure similarity to the original text, which aligns with the compression objective, and Section 4.1.1 explicitly avoids task benchmarks by saying 'existing benchmarks such as SQuAD and Piqa are mostly single-round question-answer datasets with short question length, which is not appropriate to evaluate our proposed method.' Consequently, the abstract's claim of 'better performance on different types of NLP tasks' is not established by the presented evidence; however, that is a validity gap, not a circular reduction. Hyperparameters are tuned on the same datasets, which can inflate scores, but this does not make any reported quantity equal to the input by construction. A score of 1 reflects these minor evaluation-alignment concerns rather than any derivation-by-definition.

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

The method introduces no new physical or conceptual entity, but it relies on four domain assumptions about encoder probabilities, embedding similarity, transferability, and percentile thresholding. Four free parameters, including the unreported alpha and k, govern the central mechanism, and the two most important ones are tuned on the evaluation sets rather than fixed by theory.

free parameters (4)
  • alpha = Not reported
    Weight in Eq. (1) balancing cosine similarity against log-probability; no value is given anywhere in the paper, and it directly controls which words are removed.
  • compression ratio p = 0.8, 0.6, 0.4
    Percentile threshold in Eq. (3) that determines how many words are deleted; Section 4.1 says it is tuned to optimize efficiency without degrading performance on the evaluation datasets.
  • context window size k = Not reported
    Window size in Eqs. (1)-(2) over which context probability and similarity are computed; never specified, though it changes every loss score.
  • lexical unit granularity = Not reported
    Section 4.1 says lexical unit granularity is tuned, but the paper never defines how words, phrases, and clauses are grouped into participle units.
assumptions (4)
  • domain assumption MLM probability p(xi | xi,k) measures how much information is lost when xi is removed.
    Eq. (1) in Section 3.2 treats the log-probability as an information measure without deriving it from information theory or validating it against a simpler baseline.
  • domain assumption Cosine similarity between embeddings of nearby words captures semantic redundancy.
    The similarity term in Eq. (1) is used to flag redundant words, but the embedding, the distance metric, and the justification are not specified.
  • domain assumption Pretrained transformer encoders transfer to arbitrary domains such as Reddit and arXiv abstracts.
    Section 4.5 assumes a 512-token BERT window is sufficient for all tested text categories; no per-domain validation or comparison against longer-context encoders is provided.
  • domain assumption Removing all units with loss above the p-th percentile preserves the information a downstream task needs.
    Eqs. (3)-(4) in Section 3.3 define the deletion rule, but no experiment measures task accuracy, so the preservation of task-relevant information is assumed rather than shown.

how reviews work

0 comments
Cite this review

Pith. "Pith review of ICPC: In-context Prompt Compression with Faster Inference." pith.science (2026). https://pith.science/paper/YWMQM5NA

@misc{pith2026250101625,
  author       = {Pith},
  title        = {Pith review of: ICPC: In-context Prompt Compression with Faster Inference},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YWMQM5NA}},
  note         = {Machine review of arXiv:2501.01625}
}
read the original abstract

Despite the recent success of Large Language Models (LLMs), it remains challenging to feed LLMs with long prompts due to the fixed size of LLM inputs. As a remedy, prompt compression becomes a promising solution by removing redundant tokens in the prompt. However, using LLM in the existing works requires additional computation resources and leads to memory overheads. To address it, we propose ICPC (In-context Prompt Compression), a novel and scalable prompt compression method that adaptively reduces the prompt length. The key idea of ICPC is to calculate the probability of each word appearing in the prompt using encoders and calculate information carried by each word through the information function, which effectively reduces the information loss during prompt compression and increases the speed of compression. Empirically, we demonstrate that ICPC can effectively compress long texts of different categories and thus achieve better performance and speed on different types of NLP tasks.

Figures

Figures reproduced from arXiv: 2501.01625 by the authors.

Figure 1
Figure 1. Texts before and after compression. Yellow represents words with higher importance. Up: text before [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

27 extracted references · 5 canonical work pages

  1. [1]

    Gpt-4 technical report

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. Gpt-4 technical report. arXiv preprint arXiv:2303.08774, 2023

  2. [2]

    Beyond efficiency: A systematic survey of resource-efficient large language models

    Guangji Bai, Zheng Chai, Chen Ling, Shiyu Wang, Jiaying Lu, Nan Zhang, Tingwei Shi, Ziyang Yu, Mengdan Zhu, Yifei Zhang, et al. Beyond efficiency: A systematic survey of resource-efficient large language models. arXiv preprint arXiv:2401.00625, 2024

  3. [3]

    Piqa: Reasoning about physical commonsense in natural language

    Yonatan Bisk, Rowan Zellers, Jianfeng Gao, Yejin Choi, et al. Piqa: Reasoning about physical commonsense in natural language. In Proceedings of the AAAI conference on artificial intelligence, volume 34, pages 7432–7439, 2020

  4. [4]

    Generating long sequences with sparse transformers

    Rewon Child, Scott Gray, Alec Radford, and Ilya Sutskever. Generating long sequences with sparse transformers. arXiv preprint arXiv:1904.10509, 2019

  5. [5]

    Bert: Pre-training of deep bidirectional transformers for language understanding

    Jacob Devlin. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805, 2018

  6. [6]

    A survey on large language models: Applications, challenges, limitations, and practical usage

    Muhammad Usman Hadi, Rizwan Qureshi, Abbas Shah, Muhammad Irfan, Anas Zafar, Muhammad Bilal Shaikh, Naveed Akhtar, Jia Wu, Seyedali Mirjalili, et al. A survey on large language models: Applications, challenges, limitations, and practical usage. Authorea Preprints, 2023

  7. [7]

    Deberta: Decoding-enhanced bert with disentan- gled attention

    Pengcheng He, Xiaodong Liu, Jianfeng Gao, and Weizhu Chen. Deberta: Decoding-enhanced bert with disentan- gled attention. arXiv preprint arXiv:2006.03654, 2020

  8. [8]

    Lora: Low-rank adaptation of large language models

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. Lora: Low-rank adaptation of large language models. arXiv preprint arXiv:2106.09685, 2021

Show all 27 references
  1. [9]

    The distribution of the flora in the alpine zone

    Paul Jaccard. The distribution of the flora in the alpine zone. 1. New phytologist, 11(2):37–50, 1912

  2. [10]

    Derivation of new readability formulas (automated readability index, fog count and flesch reading ease formula) for navy enlisted personnel

    JP Kincaid. Derivation of new readability formulas (automated readability index, fog count and flesch reading ease formula) for navy enlisted personnel. Chief of Naval Technical Training, 1975

  3. [11]

    Albert: A lite bert for self-supervised learning of language representations

    Z Lan. Albert: A lite bert for self-supervised learning of language representations. arXiv preprint arXiv:1909.11942, 2019

  4. [12]

    Compressing context to enhance inference efficiency of large language models

    Yucheng Li, Bo Dong, Chenghua Lin, and Frank Guerin. Compressing context to enhance inference efficiency of large language models. arXiv preprint arXiv:2310.06201, 2023

  5. [13]

    Prompt compression for large language models: A survey

    Zongqian Li, Yinhong Liu, Yixuan Su, and Nigel Collier. Prompt compression for large language models: A survey. arXiv preprint arXiv:2410.12388, 2024

  6. [14]

    Rouge: A package for automatic evaluation of summaries

    Chin-Yew Lin. Rouge: A package for automatic evaluation of summaries. In Text summarization branches out, pages 74–81, 2004

  7. [15]

    Roberta: A robustly optimized bert pretraining approach

    Yinhan Liu. Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692, 364, 2019

  8. [16]

    Large language models: A survey

    Shervin Minaee, Tomas Mikolov, Narjes Nikzad, Meysam Chenaghlu, Richard Socher, Xavier Amatriain, and Jianfeng Gao. Large language models: A survey. arXiv preprint arXiv:2402.06196, 2024

  9. [17]

    Bleu: a method for automatic evaluation of machine translation

    Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. Bleu: a method for automatic evaluation of machine translation. In Proceedings of the 40th annual meeting of the Association for Computational Linguistics, pages 311–318, 2002

  10. [18]

    Exploring the limits of transfer learning with a unified text-to-text transformer

    Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J Liu. Exploring the limits of transfer learning with a unified text-to-text transformer. Journal of machine learning research, 21(140):1–67, 2020

  11. [19]

    Know what you don’t know: Unanswerable questions for squad

    Pranav Rajpurkar, Robin Jia, and Percy Liang. Know what you don’t know: Unanswerable questions for squad. arXiv preprint arXiv:1806.03822, 2018

  12. [20]

    Benefits and challenges of variable compression ratio (vcr)

    Martyn Roberts. Benefits and challenges of variable compression ratio (vcr). Technical report, SAE Technical Paper, 2003

  13. [21]

    A mathematical theory of communication

    Claude Elwood Shannon. A mathematical theory of communication. The Bell system technical journal, 27(3):379– 423, 1948

  14. [22]

    A statistical interpretation of term specificity and its application in retrieval

    Karen Sparck Jones. A statistical interpretation of term specificity and its application in retrieval. Journal of documentation, 28(1):11–21, 1972. 7 ICPC: In-context Prompt Compression with Faster Inference A PREPRINT

  15. [23]

    Llama 2: Open foundation and fine-tuned chat models

    Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, et al. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288, 2023

  16. [24]

    Xlnet: Generalized autoregressive pretraining for language understanding

    Zhilin Yang. Xlnet: Generalized autoregressive pretraining for language understanding. arXiv preprint arXiv:1906.08237, 2019

  17. [25]

    Bertscore: Evaluating text generation with bert

    Tianyi Zhang, Varsha Kishore, Felix Wu, Kilian Q Weinberger, and Yoav Artzi. Bertscore: Evaluating text generation with bert. arXiv preprint arXiv:1904.09675, 2019

  18. [26]

    A survey of large language models

    Wayne Xin Zhao, Kun Zhou, Junyi Li, Tianyi Tang, Xiaolei Wang, Yupeng Hou, Yingqian Min, Beichen Zhang, Junjie Zhang, Zican Dong, et al. A survey of large language models. arXiv preprint arXiv:2303.18223, 2023

  19. [27]

    A survey on model compression for large language models

    Xunyu Zhu, Jian Li, Yong Liu, Can Ma, and Weiping Wang. A survey on model compression for large language models. Transactions of the Association for Computational Linguistics, 12:1556–1577, 2024. 8 ICPC: In-context Prompt Compression with Faster Inference A PREPRINT A Appendix...

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.