---
title: "FBS: Modeling Native Parallel Reading inside a Transformer (2026)"
description: "The author builds this behaviour into the model as the Fovea-Block-Skip Transformer (FBS). 89 references, cited by 10 papers here."
authors: ["Tongxi Wang"]
year: 2026
venue: "Findings of the Association for Computational Linguistics: ACL 2026"
source: "https://aclanthology.org/2026.findings-acl.200/"
doi: "10.18653/v1/2026.findings-acl.200"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/wang-2026-fbs-modeling-native-parallel-reading/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/wang-2026-fbs-modeling-native-parallel-reading/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/) / FBS: Modeling Native Parallel Reading inside a Transformer

# FBS: Modeling Native Parallel Reading inside a Transformer

**Authors:** Tongxi Wang · **Year:** 2026 · **Venue:** [Findings of the Association for Computational Linguistics: ACL 2026](https://md.research.elamur.ai/free/en/venues/findings/) ([Findings of the Association for Computational Linguistics 2026](https://md.research.elamur.ai/free/en/venues/findings/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.findings-acl.200/) · [PDF](https://aclanthology.org/2026.findings-acl.200.pdf) · [DOI 10.18653/v1/2026.findings-acl.200](https://doi.org/10.18653/v1/2026.findings-acl.200) · [OpenAlex](https://openalex.org/W7166806549)

**Citations:** 4 citations (OpenAlex, 2026-10-14) · [10](https://md.research.elamur.ai/free/en/papers/wang-2026-fbs-modeling-native-parallel-reading/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
Most LLMs still decode strictly one token at a time and spend about the same compute at every step. Fluent human readers work differently: they preview upcoming words, group text into phrases, and skim predictable stretches. The author builds this behaviour into the model as the Fovea-Block-Skip Transformer (FBS). Each causal layer gets three lightweight modules:

- **Parafovea-Attention Window (PAW)** forms a preview vector from the model's own multi-step next-token predictions, with a span that adapts to content.
- **Chunk-Head (CH)** detects BIOS-style chunk boundaries online and caches pooled chunk states.
- **Skip-Gate (SG)** decides, per token and layer, whether to bypass the whole block. It is trained with a straight-through estimator and a soft mixture so that hard skipping at inference matches training, and PPO tuning is optional.

FBS is obtained by continual pre-training of Qwen3-4B-Instruct on mixed Chinese–English data. It is evaluated on reasoning, knowledge, math and code benchmarks. At equal parameter count it improves quality while reducing executed compute, and it cuts wall-clock latency by roughly 30% in a 512→128 generation setting. The gains come from a mechanism inside the model rather than from an external decoding add-on.

## Key points
- Three systems share the same 4.0B parameter budget: the Qwen3-4B-Instruct baseline, FBS-S1 (PAW+CH only), and FBS-Full (PAW+CH+SG+RL). New module weights are randomly initialized, while the original attention and FFN weights are copied from the baseline.
- The comparisons are organized into groups:
  - decoding accelerators: Speculative Decoding, EAGLE-2, Medusa, Lookahead Decoding, timed with a Spec-Bench-style protocol;
  - layer/FFN skipping: FlexiDepth, FFN-SkipLLM;
  - long-context/cache methods: H2O, StreamingLLM, SnapKV, SlimInfer;
  - structural modeling: Segatron.
- Skip heatmaps show that middle layers (about 10–20) are bypassed most often. Early layers (1–4) and final layers (about 28–32) are mostly executed.
- The CH analysis compares the baseline, FBS without CH, and full FBS:

  | Metric | Baseline | FBS w/o CH | Full FBS |
  |---|---|---|---|
  | BIOS boundary F1 | 0.80 | 0.82 | 0.88 |
  | Idiom accuracy | 71% | 74% | 80% |
  | FactScore | 0.74 | 0.75 | 0.78 |

  Full FBS scores highest on all three metrics, a gain the paper considers especially relevant for idiom-dense Chinese.
- The author lists several limitations:
  - coherence can drift in very long outputs;
  - multi-step reasoning is fragile under aggressive skipping;
  - training and evaluation cover mainly Chinese–English;
  - the gate relies on likelihood-based proxies;
  - speed-ups depend on hardware;
  - aggressive settings are discouraged for code, JSON, formal math and high-stakes use.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

📚 Shelves (5): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 16,039 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 15,891 works · [🔌 Hardware and Architecture](https://md.research.elamur.ai/free/en/shelves/computer-science/hardware-and-architecture/) · 81 works

## Cite

Cite as: Tongxi Wang. 2026. FBS: Modeling Native Parallel Reading inside a Transformer. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{wang2026fbs,
  title = {{FBS: Modeling Native Parallel Reading inside a Transformer}},
  author = {Tongxi Wang},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.200},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.200}
}
```

- [← previous You Never Know a Person, You Only Know Their Defenses: Detecting Levels of Psychological Defense Mechanisms in Supportive Conversations](https://md.research.elamur.ai/free/en/papers/na-2026-you-never-know-person-you/)
- [next → Jais 2: A Family of Arabic-Centric Open Large Language Models](https://md.research.elamur.ai/free/en/papers/anwar-2026-jais-2-family-arabic-centric/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/)

## References

89 references, 29 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Jacob Austin, Augustus Odena, Maxwell I. Nye, Maarten Bosma, Henryk Michalewski, David Dohan, Ellen Jiang, Carrie J. Cai, Michael Terry, Quoc V. Le, and Charles Sutton. 2021. [Program synthesis with large language models](https://md.research.elamur.ai/free/en/papers/austin-2021-program-synthesis-large-language-models/). CoRR, abs/2108.07732.
2. He Bai, Peng Shi, Jimmy Lin, Yuqing Xie, Luchen Tan, Kun Xiong, Wen Gao, and Ming Li. 2021. Segatron: Segment-aware transformer for language modeling and understanding. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, pages 12526–12534. [find](https://scholar.google.com/scholar?q=He+Bai%2C+Peng+Shi%2C+Jimmy+Lin%2C+Yuqing+Xie%2C+Luchen+Tan%2C+Kun+Xiong%2C+Wen+Gao%2C+and+Ming+Li%2E+2021%2E+Segatron%3A+Segment%2Daware+transformer+for+language+modeling+and+understanding%2E+In+Proceedings+of+the+AAAI+Conf)
3. Christina Bazant-Kimmel. 2018. Learning to read authentic texts in chinese as a foreign language: An action research-based investigation of a new approach towards raising students’ awareness of literary function words. Vienna Journal of East Asian Studies, 10:211 – 232. [find](https://scholar.google.com/scholar?q=Christina+Bazant%2DKimmel%2E+2018%2E+Learning+to+read+authentic+texts+in+chinese+as+a+foreign+language%3A+An+action+research%2Dbased+investigation+of+a+new+approach+towards+raising+students%E2%80%99+awareness+of+litera)
4. Iz Beltagy, Matthew E. Peters, and Arman Cohan. 2020. [Longformer: The long-document transformer](https://md.research.elamur.ai/free/en/papers/beltagy-2020-longformer-long-document-transformer/). arXiv preprint arXiv:2004.05150.
5. Tianle Cai, Yuhong Li, Zhengyang Geng, Hongwu Peng, Jason D Lee, Deming Chen, and Tri Dao. 2024. [Medusa: Simple llm inference acceleration framework with multiple decoding heads](https://md.research.elamur.ai/free/en/papers/cai-2024-medusa-simple-llm-inference-acceleration/). arXiv preprint arXiv:2401.10774.
6. Charlie Chen, Sebastian Borgeaud, Geoffrey Irving, Jean-Baptiste Lespiau, Laurent Sifre, and John Jumper. 2023. Accelerating large language model decoding with speculative sampling. arXiv preprint arXiv:2302.01318. [find](https://scholar.google.com/scholar?q=Charlie+Chen%2C+Sebastian+Borgeaud%2C+Geoffrey+Irving%2C+Jean%2DBaptiste+Lespiau%2C+Laurent+Sifre%2C+and+John+Jumper%2E+2023%2E+Accelerating+large+language+model+decoding+with+speculative+sampling%2E+arXiv+preprint+arX)
7. Yao Chen, Yilong Chen, Yinqi Yang, Junyuan Shang, Zhenyu Zhang, Zefeng Zhang, Shuaiyi Nie, Shuohuan Wang, Yu Sun, Hua Wu, HaiFeng Wang, and Tingwen Liu. 2026a. Sparse growing transformer: Training-time sparse depth allocation via progressive attention looping. arXiv preprint arXiv:2603.23998. [find](https://scholar.google.com/scholar?q=Yao+Chen%2C+Yilong+Chen%2C+Yinqi+Yang%2C+Junyuan+Shang%2C+Zhenyu+Zhang%2C+Zefeng+Zhang%2C+Shuaiyi+Nie%2C+Shuohuan+Wang%2C+Yu+Sun%2C+Hua+Wu%2C+HaiFeng+Wang%2C+and+Tingwen+Liu%2E+2026a%2E+Sparse+growing+transformer%3A+Training%2Dtim)
8. Zhiwei Chen, Yupeng Hu, Zhiheng Fu, Zixu Li, Jiale Huang, Qinlei Huang, and Yinwei Wei. 2026b. Intent: Invariance and discrimination-aware noise mitigation for robust composed image retrieval. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 40, pages 20463–20471. [find](https://scholar.google.com/scholar?q=Zhiwei+Chen%2C+Yupeng+Hu%2C+Zhiheng+Fu%2C+Zixu+Li%2C+Jiale+Huang%2C+Qinlei+Huang%2C+and+Yinwei+Wei%2E+2026b%2E+Intent%3A+Invariance+and+discrimination%2Daware+noise+mitigation+for+robust+composed+image+retrieval%2E+In+Proc)
9. Zhiwei Chen, Yupeng Hu, Zixu Li, Zhiheng Fu, Xuemeng Song, and Liqiang Nie. 2025. Offset: Segmentation-based focus shift revision for composed image retrieval. In Proceedings of the ACM International Conference on Multimedia, page 6113–6122. [find](https://scholar.google.com/scholar?q=Zhiwei+Chen%2C+Yupeng+Hu%2C+Zixu+Li%2C+Zhiheng+Fu%2C+Xuemeng+Song%2C+and+Liqiang+Nie%2E+2025%2E+Offset%3A+Segmentation%2Dbased+focus+shift+revision+for+composed+image+retrieval%2E+In+Proceedings+of+the+ACM+International+)
10. Chung-Cheng Chiu and Colin Raffel. 2018. Monotonic chunkwise attention. In International Conference on Learning Representations. [find](https://scholar.google.com/scholar?q=Chung%2DCheng+Chiu+and+Colin+Raffel%2E+2018%2E+Monotonic+chunkwise+attention%2E+In+International+Conference+on+Learning+Representations%2E)
11. Karl Cobbe, Vineet Kosaraju, Mohammad Bavarian, Mark Chen, Heewoo Jun, Lukasz Kaiser, Matthias Plappert, Jerry Tworek, Jacob Hilton, Reiichiro Nakano, Christopher Hesse, and John Schulman. 2021. [Training verifiers to solve math word problems](https://md.research.elamur.ai/free/en/papers/cobbe-2021-training-verifiers-solve-math-word/). CoRR, abs/2110.14168.
12. Zihang Dai, Zhilin Yang, Yiming Yang, Jaime G Carbonell, Quoc Le, and Ruslan Salakhutdinov. 2019. Transformer-xl: Attentive language models beyond a fixed-length context. In Proceedings of the 57th annual meeting of the association for computational linguistics, pages 2978–2988. [find](https://scholar.google.com/scholar?q=Zihang+Dai%2C+Zhilin+Yang%2C+Yiming+Yang%2C+Jaime+G+Carbonell%2C+Quoc+Le%2C+and+Ruslan+Salakhutdinov%2E+2019%2E+Transformer%2Dxl%3A+Attentive+language+models+beyond+a+fixed%2Dlength+context%2E+In+Proceedings+of+the+57th+an)
13. Tri Dao, Daniel Y. Fu, Stefano Ermon, Atri Rudra, and Christopher Ré. 2022. Flashattention: Fast and memory-efficient exact attention with IO-awareness. In Advances in Neural Information Processing Systems. [find](https://scholar.google.com/scholar?q=Tri+Dao%2C+Daniel+Y%2E+Fu%2C+Stefano+Ermon%2C+Atri+Rudra%2C+and+Christopher+R%C3%A9%2E+2022%2E+Flashattention%3A+Fast+and+memory%2Defficient+exact+attention+with+IO%2Dawareness%2E+In+Advances+in+Neural+Information+Processing+Sy)
14. Haonan Dong, Kehan Jiang, Haoran Ye, Wenhao Zhu, Zhaolu Kang, and Guojie Song. 2026. [Neureasoner: Towards explainable, controllable, and unified reasoning via mixture-of-neurons](https://md.research.elamur.ai/free/en/papers/dong-2026-neureasoner-explainable-controllable-unified-reasoning/). Preprint, arXiv:2604.02972.
15. Haonan Dong, Wenhao Zhu, Guojie Song, and Liang Wang. 2025. AuroRA: Breaking low-rank bottleneck of loRA with nonlinear mapping. In The Thirty-ninth Annual Conference on Neural Information Processing Systems. [find](https://scholar.google.com/scholar?q=Haonan+Dong%2C+Wenhao+Zhu%2C+Guojie+Song%2C+and+Liang+Wang%2E+2025%2E+AuroRA%3A+Breaking+low%2Drank+bottleneck+of+loRA+with+nonlinear+mapping%2E+In+The+Thirty%2Dninth+Annual+Conference+on+Neural+Information+Processing+)
16. Geoffrey B. Duggan and S. Payne. 2009. Text skimming: the process and effectiveness of foraging through text under time pressure. Journal of experimental psychology. Applied, 15 3:228–42. [find](https://scholar.google.com/scholar?q=Geoffrey+B%2E+Duggan+and+S%2E+Payne%2E+2009%2E+Text+skimming%3A+the+process+and+effectiveness+of+foraging+through+text+under+time+pressure%2E+Journal+of+experimental+psychology%2E+Applied%2C+15+3%3A228%E2%80%9342%2E)
17. Mostafa Elhoushi, Rui Zhang, Md. Rizwan Islam, Ines Bouaziz, and Ming-Wei Chang. 2024. [Layerskip: Enabling early exit inference and self-speculative decoding](https://md.research.elamur.ai/free/en/papers/elhoushi-2024-layerskip-enabling-early-exit-inference/). In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 12622–12642.
18. Ralf Engbert, Antje Nuthmann, Eike M. Richter, and Reinhold Kliegl. 2005. SWIFT: A dynamical model of saccade generation during reading. Psychological Review, 112(4):777–813. [find](https://scholar.google.com/scholar?q=Ralf+Engbert%2C+Antje+Nuthmann%2C+Eike+M%2E+Richter%2C+and+Reinhold+Kliegl%2E+2005%2E+SWIFT%3A+A+dynamical+model+of+saccade+generation+during+reading%2E+Psychological+Review%2C+112%284%29%3A777%E2%80%93813%2E)
19. Angela Fan, Edouard Grave, and Armand Joulin. 2019. [Reducing transformer depth on demand with structured dropout](https://md.research.elamur.ai/free/en/papers/fan-2019-reducing-transformer-depth-demand-structured/). arXiv preprint arXiv:1909.11556.
20. Gemma Fitzsimmons, M. Weal, and Denis Drieghe. 2014. Skim reading: an adaptive strategy for reading on the web. In Web Science Conference. [find](https://scholar.google.com/scholar?q=Gemma+Fitzsimmons%2C+M%2E+Weal%2C+and+Denis+Drieghe%2E+2014%2E+Skim+reading%3A+an+adaptive+strategy+for+reading+on+the+web%2E+In+Web+Science+Conference%2E)
21. Karl Friston. 2005. A theory of cortical responses. Philosophical Transactions of the Royal Society B: Biological Sciences, 360(1456):815–836. [find](https://scholar.google.com/scholar?q=Karl+Friston%2E+2005%2E+A+theory+of+cortical+responses%2E+Philosophical+Transactions+of+the+Royal+Society+B%3A+Biological+Sciences%2C+360%281456%29%3A815%E2%80%93836%2E)
22. Itai Gat, Heli Ben-Hamu, Marton Havasi, Daniel Haziza, Jeremy Reizenstein, Gabriel Synnaeve, David Lopez-Paz, Brian Karrer, and Yaron Lipman. 2025. Set block decoding is a language model inference accelerator. arXiv preprint arXiv:2509.04185. — [arXiv:2509.04185](https://arxiv.org/abs/2509.04185) [find](https://scholar.google.com/scholar?q=Itai+Gat%2C+Heli+Ben%2DHamu%2C+Marton+Havasi%2C+Daniel+Haziza%2C+Jeremy+Reizenstein%2C+Gabriel+Synnaeve%2C+David+Lopez%2DPaz%2C+Brian+Karrer%2C+and+Yaron+Lipman%2E+2025%2E+Set+block+decoding+is+a+language+model+inference+acc)
23. Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Peiyi Wang, Qihao Zhu, Runxin Xu, Ruoyu Zhang, Shirong Ma, Xiao Bi, Xiaokang Zhang, Xingkai Yu, Yu Wu, Z. F. Wu, Zhibin Gou, Zhihong Shao, Zhuoshu Li, Ziyi Gao, Aixin Liu, and 175 others. 2025. Deepseek-r1 incentivizes reasoning in llms through rein… [find](https://scholar.google.com/scholar?q=Daya+Guo%2C+Dejian+Yang%2C+Haowei+Zhang%2C+Junxiao+Song%2C+Peiyi+Wang%2C+Qihao+Zhu%2C+Runxin+Xu%2C+Ruoyu+Zhang%2C+Shirong+Ma%2C+Xiao+Bi%2C+Xiaokang+Zhang%2C+Xingkai+Yu%2C+Yu+Wu%2C+Z%2E+F%2E+Wu%2C+Zhibin+Gou%2C+Zhihong+Shao%2C+Zhuoshu+Li)
24. Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika, Dawn Song, and Jacob Steinhardt. 2021. [Measuring massive multitask language understanding](https://md.research.elamur.ai/free/en/papers/hendrycks-2020-measuring-massive-multitask-language-understanding/). Preprint, arXiv:2009.03300.
25. Yupeng Hu, Zixu Li, Zhiwei Chen, Qinlei Huang, Zhiheng Fu, Mingzhu Xu, and Liqiang Nie. 2026. Refine: Composed video retrieval via shared and differential semantics enhancement. ACM Transactions on Multimedia Computing, Communications and Applications. [find](https://scholar.google.com/scholar?q=Yupeng+Hu%2C+Zixu+Li%2C+Zhiwei+Chen%2C+Qinlei+Huang%2C+Zhiheng+Fu%2C+Mingzhu+Xu%2C+and+Liqiang+Nie%2E+2026%2E+Refine%3A+Composed+video+retrieval+via+shared+and+differential+semantics+enhancement%2E+ACM+Transactions+on+Mu)
26. Yuzhen Huang, Yuzhuo Bai, Zhihao Zhu, Junlei Zhang, Jinghan Zhang, Tangjun Su, Junteng Liu, Chuancheng Lv, Yikai Zhang, Jiayi Lei, Yao Fu, Maosong Sun, and Junxian He. 2023. [C-eval: A multi-level multi-discipline chinese evaluation suite for foundation models](https://md.research.elamur.ai/free/en/papers/huang-2023-c-eval-multi-level-multi/).
27. Ajay Kumar Jaiswal, Bodun Hu, Lu Yin, Yeonju Ro, Tianlong Chen, Shiwei Liu, and Aditya Akella. 2024. Ffn-skipllm: A hidden gem for autoregressive decoding with adaptive feed forward skipping. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 16943–16956… [find](https://scholar.google.com/scholar?q=Ajay+Kumar+Jaiswal%2C+Bodun+Hu%2C+Lu+Yin%2C+Yeonju+Ro%2C+Tianlong+Chen%2C+Shiwei+Liu%2C+and+Aditya+Akella%2E+2024%2E+Ffn%2Dskipllm%3A+A+hidden+gem+for+autoregressive+decoding+with+adaptive+feed+forward+skipping%2E+In+Proce)
28. Tim Jansen, Yangling Tong, Victoria Zevallos, and Pedro Ortiz Suarez. 2022. Perplexed by quality: A perplexity-based method for adult and harmful content detection in multilingual heterogeneous web data. arXiv preprint arXiv:2212.10440. [find](https://scholar.google.com/scholar?q=Tim+Jansen%2C+Yangling+Tong%2C+Victoria+Zevallos%2C+and+Pedro+Ortiz+Suarez%2E+2022%2E+Perplexed+by+quality%3A+A+perplexity%2Dbased+method+for+adult+and+harmful+content+detection+in+multilingual+heterogeneous+web+da)
29. Kehan Jiang, Haonan Dong, Zhaolu Kang, Zhengzhou Zhu, and Guojie Song. 2026. [Foe: Forest of errors makes the first solution the best in large reasoning models](https://md.research.elamur.ai/free/en/papers/jiang-2026-foe-forest-errors-makes-first/). Preprint, arXiv:2604.02967.
30. Aditya Jonnalagadda, William Yang Wang, BS Manjunath, and Miguel P Eckstein. 2021. Foveater: Foveated transformer for image classification. arXiv preprint arXiv:2105.14173. — [arXiv:2105.14173](https://arxiv.org/abs/2105.14173) [find](https://scholar.google.com/scholar?q=Aditya+Jonnalagadda%2C+William+Yang+Wang%2C+BS+Manjunath%2C+and+Miguel+P+Eckstein%2E+2021%2E+Foveater%3A+Foveated+transformer+for+image+classification%2E+arXiv+preprint+arXiv%3A2105%2E14173%2E+%E2%80%94+%5BarXiv%3A2105%2E14173%5D%28https%3A)
31. Qizhen Lan, Jung Im Choi, and Qing Tian. 2026a. Visual detector compression via location-aware discriminant analysis. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 3546–3555. [find](https://scholar.google.com/scholar?q=Qizhen+Lan%2C+Jung+Im+Choi%2C+and+Qing+Tian%2E+2026a%2E+Visual+detector+compression+via+location%2Daware+discriminant+analysis%2E+In+Proceedings+of+the+IEEE%2FCVF+Winter+Conference+on+Applications+of+Computer+Visio)
32. Qizhen Lan, Yu-Chun Hsu, Nida Saddaf Khan, and Xiaoqian Jiang. 2026b. Reco-kd: Region-and context-aware knowledge distillation for efficient 3d medical image segmentation. arXiv preprint arXiv:2601.08301. — [arXiv:2601.08301](https://arxiv.org/abs/2601.08301) [find](https://scholar.google.com/scholar?q=Qizhen+Lan%2C+Yu%2DChun+Hsu%2C+Nida+Saddaf+Khan%2C+and+Xiaoqian+Jiang%2E+2026b%2E+Reco%2Dkd%3A+Region%2Dand+context%2Daware+knowledge+distillation+for+efficient+3d+medical+image+segmentation%2E+arXiv+preprint+arXiv%3A2601%2E08)
33. Yaniv Leviathan, Matan Kalman, and Yossi Matias. 2023. Fast inference from transformers via speculative decoding. In International Conference on Machine Learning, pages 19274–19286. PMLR. [find](https://scholar.google.com/scholar?q=Yaniv+Leviathan%2C+Matan+Kalman%2C+and+Yossi+Matias%2E+2023%2E+Fast+inference+from+transformers+via+speculative+decoding%2E+In+International+Conference+on+Machine+Learning%2C+pages+19274%E2%80%9319286%2E+PMLR%2E)
34. Guocong Li, Jinjian Zhang, Ping Wang, Dongnan Liu, Tian Liang, Qiuyi Qi, Hao Huang, Siyan Guo, Mutian Bao, Wei Zhou, Linjian Mo, Hongxia Xu, and Jian Wu. 2026a. Mol: Adaptive mixture-of-length reasoning for efficient question answering with context. In The Fourteenth International Conference on Lear… [find](https://scholar.google.com/scholar?q=Guocong+Li%2C+Jinjian+Zhang%2C+Ping+Wang%2C+Dongnan+Liu%2C+Tian+Liang%2C+Qiuyi+Qi%2C+Hao+Huang%2C+Siyan+Guo%2C+Mutian+Bao%2C+Wei+Zhou%2C+Linjian+Mo%2C+Hongxia+Xu%2C+and+Jian+Wu%2E+2026a%2E+Mol%3A+Adaptive+mixture%2Dof%2Dlength+reasoni)
35. Haonan Li, Yixuan Zhang, Fajri Koto, Yifei Yang, Hai Zhao, Yeyun Gong, Nan Duan, and Timothy Baldwin. 2024a. [Cmmlu: Measuring massive multitask language understanding in chinese](https://md.research.elamur.ai/free/en/papers/li-2023-cmmlu-measuring-massive-multitask-language/). In Findings of the Association for Computational Linguistics: ACL 2024, pages 11260–11285.
36. Yuhong Li, Yingbing Huang, Bowen Yang, Bharat Venkitesh, Acyr Locatelli, Hanchen Ye, Tianle Cai, Patrick Lewis, and Deming Chen. 2024b. [Snapkv: Llm knows what you are looking for before generation](https://md.research.elamur.ai/free/en/papers/li-2024-snapkv-llm-knows-looking-before/). Advances in Neural Information Processing Systems, 37:22947–22970.
37. Yuhui Li, Fangyun Wei, Chao Zhang, and Hongyang Zhang. 2024c. [Eagle-2: Faster inference of language models with dynamic draft trees](https://md.research.elamur.ai/free/en/papers/li-2024-eagle-2-faster-inference-language/). In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 7421–7432.
38. Yuhui Li, Fangyun Wei, Chao Zhang, and Hongyang Zhang. 2024d. [Eagle: speculative sampling requires rethinking feature uncertainty](https://md.research.elamur.ai/free/en/papers/li-2024-eagle-speculative-sampling-requires-rethinking/). In Proceedings of the 41st International Conference on Machine Learning, pages 28935–28948.
39. Zixu Li, Yupeng Hu, Zhiwei Chen, Qinlei Huang, Guozhi Qiu, Zhiheng Fu, and Meng Liu. 2026b. Retrack: Evidence-driven dual-stream directional anchor calibration network for composed video retrieval. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 40, pages 23373–23381. [find](https://scholar.google.com/scholar?q=Zixu+Li%2C+Yupeng+Hu%2C+Zhiwei+Chen%2C+Qinlei+Huang%2C+Guozhi+Qiu%2C+Zhiheng+Fu%2C+and+Meng+Liu%2E+2026b%2E+Retrack%3A+Evidence%2Ddriven+dual%2Dstream+directional+anchor+calibration+network+for+composed+video+retrieval%2E+In)
40. Zixu Li, Yupeng Hu, Zhiwei Chen, Shiqi Zhang, Qinlei Huang, Zhiheng Fu, and Yinwei Wei. 2026c. Habit: Chrono-synergia robust progressive learning framework for composed image retrieval. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 40, pages 6762–6770. [find](https://scholar.google.com/scholar?q=Zixu+Li%2C+Yupeng+Hu%2C+Zhiwei+Chen%2C+Shiqi+Zhang%2C+Qinlei+Huang%2C+Zhiheng+Fu%2C+and+Yinwei+Wei%2E+2026c%2E+Habit%3A+Chrono%2Dsynergia+robust+progressive+learning+framework+for+composed+image+retrieval%2E+In+Proceedings)
41. Yule Liu, Jingyi Zheng, Zhen Sun, Zifan Peng, Wenhan Dong, Zeyang Sha, Shiwen Cui, Weiqiang Wang, and Xinlei He. 2025. Thought manipulation: External thought can be efficient for large reasoning models. arXiv preprint arXiv:2504.13626. — [arXiv:2504.13626](https://arxiv.org/abs/2504.13626) [find](https://scholar.google.com/scholar?q=Yule+Liu%2C+Jingyi+Zheng%2C+Zhen+Sun%2C+Zifan+Peng%2C+Wenhan+Dong%2C+Zeyang+Sha%2C+Shiwen+Cui%2C+Weiqiang+Wang%2C+and+Xinlei+He%2E+2025%2E+Thought+manipulation%3A+External+thought+can+be+efficient+for+large+reasoning+model)
42. Lingkun Long, Rubing Yang, Yushi Huang, Desheng Hui, Ao Zhou, and Jianlei Yang. 2025. Sliminfer: Accelerating long-context llm inference via dynamic token pruning. arXiv preprint arXiv:2508.06447. [find](https://scholar.google.com/scholar?q=Lingkun+Long%2C+Rubing+Yang%2C+Yushi+Huang%2C+Desheng+Hui%2C+Ao+Zhou%2C+and+Jianlei+Yang%2E+2025%2E+Sliminfer%3A+Accelerating+long%2Dcontext+llm+inference+via+dynamic+token+pruning%2E+arXiv+preprint+arXiv%3A2508%2E06447%2E)
43. Ruixuan Luo, Jingjing Xu, Yi Zhang, Zhiyuan Zhang, Xuancheng Ren, and Xu Sun. 2019. Pkuseg: A toolkit for multi-domain chinese word segmentation. arXiv preprint arXiv:1906.11455. — [arXiv:1906.11455](https://arxiv.org/abs/1906.11455) [find](https://scholar.google.com/scholar?q=Ruixuan+Luo%2C+Jingjing+Xu%2C+Yi+Zhang%2C+Zhiyuan+Zhang%2C+Xuancheng+Ren%2C+and+Xu+Sun%2E+2019%2E+Pkuseg%3A+A+toolkit+for+multi%2Ddomain+chinese+word+segmentation%2E+arXiv+preprint+arXiv%3A1906%2E11455%2E+%E2%80%94+%5BarXiv%3A1906%2E11455%5D%28)
44. Xuan Luo, Weizhi Wang, and Xifeng Yan. 2025. Adaptive layer-skipping in pre-trained llms. arXiv preprint arXiv:2503.23798. — [arXiv:2503.23798](https://arxiv.org/abs/2503.23798) [find](https://scholar.google.com/scholar?q=Xuan+Luo%2C+Weizhi+Wang%2C+and+Xifeng+Yan%2E+2025%2E+Adaptive+layer%2Dskipping+in+pre%2Dtrained+llms%2E+arXiv+preprint+arXiv%3A2503%2E23798%2E+%E2%80%94+%5BarXiv%3A2503%2E23798%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F2503%2E23798%29)
45. Mingbo Ma, Liang Huang, Hao Xiong, Renjie Zheng, Kaibo Liu, Baigong Zheng, Chuanqiang Zhang, Zhongjun He, Hairong Liu, Xing Li, Hua Wu, and Haifeng Wang. 2019. [STACL: Simultaneous translation with implicit anticipation and controllable latency using prefix-to-prefix framework](https://md.research.elamur.ai/free/en/papers/ma-2019-stacl-simultaneous-translation-implicit-anticipation/). In Proceedings of the…
46. Juhong Min, Yucheng Zhao, Chong Luo, and Minsu Cho. 2022. Peripheral vision transformer. In Advances in Neural Information Processing Systems. [find](https://scholar.google.com/scholar?q=Juhong+Min%2C+Yucheng+Zhao%2C+Chong+Luo%2C+and+Minsu+Cho%2E+2022%2E+Peripheral+vision+transformer%2E+In+Advances+in+Neural+Information+Processing+Systems%2E)
47. Shervin Minaee, Tomáš Mikolov, Narjes Nikzad, M. Chenaghlu, R. Socher, Xavier Amatriain, and Jianfeng Gao. 2024. [Large language models: A survey](https://md.research.elamur.ai/free/en/papers/minaee-2024-large-language-models-survey/). ArXiv, abs/2402.06196.
48. Guozhi Qiu, Zhiwei Chen, Zixu Li, Qinlei Huang, Zhiheng Fu, Xuemeng Song, and Yupeng Hu. 2026. Melt: Improve composed image retrieval via the modification frequentation-rarity balance network. arXiv preprint arXiv:2603.29291. — [arXiv:2603.29291](https://arxiv.org/abs/2603.29291) [find](https://scholar.google.com/scholar?q=Guozhi+Qiu%2C+Zhiwei+Chen%2C+Zixu+Li%2C+Qinlei+Huang%2C+Zhiheng+Fu%2C+Xuemeng+Song%2C+and+Yupeng+Hu%2E+2026%2E+Melt%3A+Improve+composed+image+retrieval+via+the+modification+frequentation%2Drarity+balance+network%2E+arXiv+p)
49. Colin Raffel, Minh-Thang Luong, Peter J. Liu, Ron J. Weiss, and Douglas Eck. 2017. Online and linear-time attention by enforcing monotonic alignments. In Proceedings of the 34th International Conference on Machine Learning, volume 70 of Proceedings of Machine Learning Research, pages 2837–2846. PMLR… [find](https://scholar.google.com/scholar?q=Colin+Raffel%2C+Minh%2DThang+Luong%2C+Peter+J%2E+Liu%2C+Ron+J%2E+Weiss%2C+and+Douglas+Eck%2E+2017%2E+Online+and+linear%2Dtime+attention+by+enforcing+monotonic+alignments%2E+In+Proceedings+of+the+34th+International+Conferen)
50. Rajesh P. N. Rao and Dana H. Ballard. 1999. Predictive coding in the visual cortex: a functional interpretation of some extra-classical receptive-field effects. Nature Neuroscience, 2(1):79–87. [find](https://scholar.google.com/scholar?q=Rajesh+P%2E+N%2E+Rao+and+Dana+H%2E+Ballard%2E+1999%2E+Predictive+coding+in+the+visual+cortex%3A+a+functional+interpretation+of+some+extra%2Dclassical+receptive%2Dfield+effects%2E+Nature+Neuroscience%2C+2%281%29%3A79%E2%80%9387%2E)
51. Keith Rayner. 1975. The perceptual span and peripheral cues in reading. Cognitive Psychology, 7(1):65–81. [find](https://scholar.google.com/scholar?q=Keith+Rayner%2E+1975%2E+The+perceptual+span+and+peripheral+cues+in+reading%2E+Cognitive+Psychology%2C+7%281%29%3A65%E2%80%9381%2E)
52. Keith Rayner. 1998. Eye movements in reading and information processing: 20 years of research. Psychological Bulletin, 124(3):372–422. [find](https://scholar.google.com/scholar?q=Keith+Rayner%2E+1998%2E+Eye+movements+in+reading+and+information+processing%3A+20+years+of+research%2E+Psychological+Bulletin%2C+124%283%29%3A372%E2%80%93422%2E)
53. Erik D. Reichle, Alexander Pollatsek, Deborah L. Fisher, and Keith Rayner. 1998. Toward a model of eye movement control in reading. Psychological Review, 105(1):125–157. [find](https://scholar.google.com/scholar?q=Erik+D%2E+Reichle%2C+Alexander+Pollatsek%2C+Deborah+L%2E+Fisher%2C+and+Keith+Rayner%2E+1998%2E+Toward+a+model+of+eye+movement+control+in+reading%2E+Psychological+Review%2C+105%281%29%3A125%E2%80%93157%2E)
54. Erik D. Reichle, Keith Rayner, and Alexander Pollatsek. 2003. The E-Z reader model of eye-movement control in reading: Comparisons to other models. Behavioral and Brain Sciences, 26(4):445–476. [find](https://scholar.google.com/scholar?q=Erik+D%2E+Reichle%2C+Keith+Rayner%2C+and+Alexander+Pollatsek%2E+2003%2E+The+E%2DZ+reader+model+of+eye%2Dmovement+control+in+reading%3A+Comparisons+to+other+models%2E+Behavioral+and+Brain+Sciences%2C+26%284%29%3A445%E2%80%93476%2E)
55. Elizabeth R Schotter. 2013. Synonyms provide semantic preview benefit in english. Journal of Memory and Language, 69(4):619–633. [find](https://scholar.google.com/scholar?q=Elizabeth+R+Schotter%2E+2013%2E+Synonyms+provide+semantic+preview+benefit+in+english%2E+Journal+of+Memory+and+Language%2C+69%284%29%3A619%E2%80%93633%2E)
56. Elizabeth R. Schotter, Bernhard Angele, and Keith Rayner. 2012. Parafoveal processing in reading. Attention, Perception, & Psychophysics, 74(1):5–35. [find](https://scholar.google.com/scholar?q=Elizabeth+R%2E+Schotter%2C+Bernhard+Angele%2C+and+Keith+Rayner%2E+2012%2E+Parafoveal+processing+in+reading%2E+Attention%2C+Perception%2C+%26+Psychophysics%2C+74%281%29%3A5%E2%80%9335%2E)
57. Yuhao Shen, Tianyu Liu, Junyi Shen, Jinyang Wu, Quan Kong, Li Huan, and Cong Wang. 2026. Double: Breaking the acceleration limit via double retrieval speculative parallelism. arXiv preprint arXiv:2601.05524. [find](https://scholar.google.com/scholar?q=Yuhao+Shen%2C+Tianyu+Liu%2C+Junyi+Shen%2C+Jinyang+Wu%2C+Quan+Kong%2C+Li+Huan%2C+and+Cong+Wang%2E+2026%2E+Double%3A+Breaking+the+acceleration+limit+via+double+retrieval+speculative+parallelism%2E+arXiv+preprint+arXiv%3A2601)
58. Yangyang Shi, Yongqiang Wang, Chunyang Wu, Ching-Feng Yeh, Julian Chan, Frank Zhang, Duc Le, and Mike Seltzer. 2021. Emformer: Efficient memory transformer based acoustic model for low latency streaming speech recognition. In 2021 IEEE International Conference on Acoustics, Speech and Signal Process… [find](https://scholar.google.com/scholar?q=Yangyang+Shi%2C+Yongqiang+Wang%2C+Chunyang+Wu%2C+Ching%2DFeng+Yeh%2C+Julian+Chan%2C+Frank+Zhang%2C+Duc+Le%2C+and+Mike+Seltzer%2E+2021%2E+Emformer%3A+Efficient+memory+transformer+based+acoustic+model+for+low+latency+streami)
59. Joshua Snell, Martijn Meeter, and Jonathan Grainger. 2017. Evidence for simultaneous syntactic processing of multiple words during reading. PloS one, 12(3):e0173720. [find](https://scholar.google.com/scholar?q=Joshua+Snell%2C+Martijn+Meeter%2C+and+Jonathan+Grainger%2E+2017%2E+Evidence+for+simultaneous+syntactic+processing+of+multiple+words+during+reading%2E+PloS+one%2C+12%283%29%3Ae0173720%2E)
60. Mirac Suzgun, Nathan Scales, Nathanael Schärli, Sebastian Gehrmann, Yi Tay, Hyung Won Chung, Aakanksha Chowdhery, Quoc V. Le, Ed H. Chi, Denny Zhou, and Jason Wei. 2023. [Challenging big-bench tasks and whether chain-of-thought can solve them](https://md.research.elamur.ai/free/en/papers/suzgun-2023-challenging-big-bench-tasks-whether/). In Findings of the Association for Computational Linguist…
61. Qwen Team. 2025. [Qwen3 technical report](https://md.research.elamur.ai/free/en/papers/yang-2025-qwen3-technical-report/). arXiv preprint arXiv:2505.09388.
62. Surat Teerapittayanon, Bradley McDanel, and Hsiang-Tsung Kung. 2016. Branchynet: Fast inference via early exiting from deep neural networks. In 2016 23rd international conference on pattern recognition (ICPR), pages 2464–2469. IEEE. [find](https://scholar.google.com/scholar?q=Surat+Teerapittayanon%2C+Bradley+McDanel%2C+and+Hsiang%2DTsung+Kung%2E+2016%2E+Branchynet%3A+Fast+inference+via+early+exiting+from+deep+neural+networks%2E+In+2016+23rd+international+conference+on+pattern+recognitio)
63. Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, M. Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, Aur’elien Rodriguez, Armand Joulin, Edouard Grave, and Guillaume Lample. 2023. [Llama: Open and efficient foundation language models](https://md.research.elamur.ai/free/en/papers/touvron-2023-llama-open-efficient-foundation-language/). ArXiv, abs/2302…
64. Chengbing Wang, Yang Zhang, Wenjie Wang, Xiaoyan Zhao, Fuli Feng, Xiangnan He, and Tat-Seng Chua. 2025. Think-while-generating: On-the-fly reasoning for personalized long-form generation. arXiv preprint arXiv:2512.06690. — [arXiv:2512.06690](https://arxiv.org/abs/2512.06690) [find](https://scholar.google.com/scholar?q=Chengbing+Wang%2C+Yang+Zhang%2C+Wenjie+Wang%2C+Xiaoyan+Zhao%2C+Fuli+Feng%2C+Xiangnan+He%2C+and+Tat%2DSeng+Chua%2E+2025%2E+Think%2Dwhile%2Dgenerating%3A+On%2Dthe%2Dfly+reasoning+for+personalized+long%2Dform+generation%2E+arXiv+prepri)
65. Chengbing Wang, Wuqiang Zheng, Yang Zhang, Fengbin Zhu, Junyi Cheng, Yi Xie, Wenjie Wang, and Fuli Feng. 2026a. [Perm: Psychology-grounded empathetic reward modeling for large language models](https://md.research.elamur.ai/free/en/papers/wang-2026-perm-psychology-grounded-empathetic-reward/). arXiv preprint arXiv:2601.10532.
66. Junxi Wang, Te Sun, Jiayi Zhu, Junxian Li, Haowen Xu, Zichen Wen, Xuming Hu, Zhiyu Li, and Linfeng Zhang. 2026b. [Streammeco: Long-term agent memory compression for efficient streaming video understanding](https://md.research.elamur.ai/free/en/papers/wang-2026-streammeco-long-term-agent-memory/). Preprint, arXiv:2604.09000.
67. Maurice Weber, Daniel Y. Fu, Quentin Anthony, Yonatan Oren, Shane Adams, Anton Alexandrov, Xiaozhong Lyu, Huu Nguyen, Xiaozhe Yao, Virginia Adams, Ben Athiwaratkun, Rahul Chalamala, Kezhen Chen, Max Ryabinin, Tri Dao, Percy Liang, Christopher Ré, Irina Rish, and Ce Zhang. 2024. [Redpajama: an open da…](https://md.research.elamur.ai/free/en/papers/weber-2024-redpajama-open-dataset-training-large/)
68. Tianwen Wei, Jian Luan, Wei Liu, Shuang Dong, and Bin Wang. 2023. [Cmath: Can your language model pass chinese elementary school math test?](https://md.research.elamur.ai/free/en/papers/wei-2023-cmath-language-model-pass/) arXiv preprint arXiv:2306.16636.
69. Shaohua Wu, Xudong Zhao, Shenling Wang, Jiangang Luo, Lingjun Li, Xi Chen, Bing Zhao, Wei Wang, Tong Yu, Rongguo Zhang, Jiahua Zhang, and Chao Wang. 2023. Yuan 2.0: A large language model with localized filtering-based attention. Preprint, arXiv:2311.15786. — [arXiv:2311.15786](https://arxiv.org/abs/2311.15786) [find](https://scholar.google.com/scholar?q=Shaohua+Wu%2C+Xudong+Zhao%2C+Shenling+Wang%2C+Jiangang+Luo%2C+Lingjun+Li%2C+Xi+Chen%2C+Bing+Zhao%2C+Wei+Wang%2C+Tong+Yu%2C+Rongguo+Zhang%2C+Jiahua+Zhang%2C+and+Chao+Wang%2E+2023%2E+Yuan+2%2E0%3A+A+large+language+model+with+localiz)
70. Guangxuan Xiao, Yuandong Tian, Beidi Chen, Song Han, and Mike Lewis. 2023. Efficient streaming language models with attention sinks. arXiv preprint arXiv:2309.17453. [find](https://scholar.google.com/scholar?q=Guangxuan+Xiao%2C+Yuandong+Tian%2C+Beidi+Chen%2C+Song+Han%2C+and+Mike+Lewis%2E+2023%2E+Efficient+streaming+language+models+with+attention+sinks%2E+arXiv+preprint+arXiv%3A2309%2E17453%2E)
71. Xi Xiao, Chenrui Ma, Yunbei Zhang, Chen Liu, Zhux-uanzi Wang, Yanshu Li, Lin Zhao, Guosheng Hu, Tianyang Wang, and Hao Xu. 2026. Not all directions matter: Toward structured and task-aware low-rank adaptation. arXiv preprint arXiv:2603.14228. — [arXiv:2603.14228](https://arxiv.org/abs/2603.14228) [find](https://scholar.google.com/scholar?q=Xi+Xiao%2C+Chenrui+Ma%2C+Yunbei+Zhang%2C+Chen+Liu%2C+Zhux%2Duanzi+Wang%2C+Yanshu+Li%2C+Lin+Zhao%2C+Guosheng+Hu%2C+Tianyang+Wang%2C+and+Hao+Xu%2E+2026%2E+Not+all+directions+matter%3A+Toward+structured+and+task%2Daware+low%2Drank+ad)
72. Xi Xiao, Yunbei Zhang, Xingjian Li, Tianyang Wang, Xiao Wang, Yuxiang Wei, Jihun Hamm, and Min Xu. 2025a. Visual instance-aware prompt tuning. In 33rd ACM International Conference on Multimedia, MM 2025, pages 2880–2889. Association for Computing Machinery, Inc. [find](https://scholar.google.com/scholar?q=Xi+Xiao%2C+Yunbei+Zhang%2C+Xingjian+Li%2C+Tianyang+Wang%2C+Xiao+Wang%2C+Yuxiang+Wei%2C+Jihun+Hamm%2C+and+Min+Xu%2E+2025a%2E+Visual+instance%2Daware+prompt+tuning%2E+In+33rd+ACM+International+Conference+on+Multimedia%2C+MM+20)
73. Xi Xiao, Yunbei Zhang, Lin Zhao, Yiyang Liu, Xiaoying Liao, Zheda Mai, Xingjian Li, Xiao Wang, Hao Xu, Jihun Hamm, Xue Lin, Min Xu, Qifan Wang, Tianyang Wang, and Cheng Han. 2025b. Prompt-based adaptation in large-scale vision models: A survey. arXiv preprint arXiv:2510.13219. — [arXiv:2510.13219](https://arxiv.org/abs/2510.13219) [find](https://scholar.google.com/scholar?q=Xi+Xiao%2C+Yunbei+Zhang%2C+Lin+Zhao%2C+Yiyang+Liu%2C+Xiaoying+Liao%2C+Zheda+Mai%2C+Xingjian+Li%2C+Xiao+Wang%2C+Hao+Xu%2C+Jihun+Hamm%2C+Xue+Lin%2C+Min+Xu%2C+Qifan+Wang%2C+Tianyang+Wang%2C+and+Cheng+Han%2E+2025b%2E+Prompt%2Dbased+adapta)
74. Jiawen Xie, Pengyu Cheng, Xiao Liang, Yong Dai, and Nan Du. 2024. Chunk, align, select: A simple long-sequence processing method for transformers. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 17438–17455. [find](https://scholar.google.com/scholar?q=Jiawen+Xie%2C+Pengyu+Cheng%2C+Xiao+Liang%2C+Yong+Dai%2C+and+Nan+Du%2E+2024%2E+Chunk%2C+align%2C+select%3A+A+simple+long%2Dsequence+processing+method+for+transformers%2E+In+Proceedings+of+the+62nd+Annual+Meeting+of+the+Asso)
75. Ji Xin, Raphael Tang, Jaejun Lee, Yaoliang Yu, and Jimmy Lin. 2020. [Deebert: Dynamic early exiting for accelerating bert inference](https://md.research.elamur.ai/free/en/papers/xin-2020-deebert-dynamic-early-exiting-accelerating/). In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 2246–2251.
76. Liling Xu, Sui Liu, Suiping Wang, Dongxia Sun, and Nan Li. 2022. Word’s predictability can modulate semantic preview effect in high-constraint sentences. Frontiers in Psychology, 13:849351. [find](https://scholar.google.com/scholar?q=Liling+Xu%2C+Sui+Liu%2C+Suiping+Wang%2C+Dongxia+Sun%2C+and+Nan+Li%2E+2022%2E+Word%E2%80%99s+predictability+can+modulate+semantic+preview+effect+in+high%2Dconstraint+sentences%2E+Frontiers+in+Psychology%2C+13%3A849351%2E)
77. Jinbiao Yang, Qing Cai, and Xing Tian. 2020. How do we segment text? two-stage chunking operation in reading. Eneuro, 7(3). [find](https://scholar.google.com/scholar?q=Jinbiao+Yang%2C+Qing+Cai%2C+and+Xing+Tian%2E+2020%2E+How+do+we+segment+text%3F+two%2Dstage+chunking+operation+in+reading%2E+Eneuro%2C+7%283%29%2E)
78. Qianyun Yang, Zhiwei Chen, Yupeng Hu, Zixu Li, Zhiheng Fu, and Liqiang Nie. 2026. Stable: Efficient hybrid nearest neighbor search via magnitude-uniformity and cardinality-robustness. arXiv preprint arXiv:2604.01617. — [arXiv:2604.01617](https://arxiv.org/abs/2604.01617) [find](https://scholar.google.com/scholar?q=Qianyun+Yang%2C+Zhiwei+Chen%2C+Yupeng+Hu%2C+Zixu+Li%2C+Zhiheng+Fu%2C+and+Liqiang+Nie%2E+2026%2E+Stable%3A+Efficient+hybrid+nearest+neighbor+search+via+magnitude%2Duniformity+and+cardinality%2Drobustness%2E+arXiv+preprint+a)
79. Lili Yu, Dániel Simig, Colin Flaherty, Armen Aghajanyan, Luke Zettlemoyer, and Mike Lewis. 2023. Megabyte: Predicting million-byte sequences with multiscale transformers. Advances in Neural Information Processing Systems, 36:78808–78823. [find](https://scholar.google.com/scholar?q=Lili+Yu%2C+D%C3%A1niel+Simig%2C+Colin+Flaherty%2C+Armen+Aghajanyan%2C+Luke+Zettlemoyer%2C+and+Mike+Lewis%2E+2023%2E+Megabyte%3A+Predicting+million%2Dbyte+sequences+with+multiscale+transformers%2E+Advances+in+Neural+Informatio)
80. Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontañón, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, and Amr Ahmed. 2020. [Big bird: Transformers for longer sequences](https://md.research.elamur.ai/free/en/papers/zaheer-2020-big-bird-transformers-longer-sequences/). In Advances in Neural Information Processing Systems.
81. Hao Zhang, Mengsi Lyu, Zhuo Chen, Xingrun Xing, Yulong Ao, and Yonghua Lin. 2025a. [Pdtrim: Targeted pruning for prefill-decode disaggregation in inference](https://md.research.elamur.ai/free/en/papers/zhang-2026-pdtrim-targeted-pruning-prefill-decode/). arXiv preprint arXiv:2509.04467.
82. Hao Zhang, Mengsi Lyu, Chenrui He, Yulong Ao, and Yonghua Lin. 2025b. [Trimtokenator: Towards adaptive visual token pruning for large multimodal models](https://md.research.elamur.ai/free/en/papers/zhang-2026-trimtokenator-adaptive-visual-token-pruning/). arXiv preprint arXiv:2509.00320.
83. Hao Zhang, Mengsi Lyu, Bo Huang, Yulong Ao, and Yonghua Lin. 2025c. Trimtokenator-lc: Towards adaptive visual token pruning for large multimodal models with long contexts. arXiv preprint arXiv:2512.22748. — [arXiv:2512.22748](https://arxiv.org/abs/2512.22748) [find](https://scholar.google.com/scholar?q=Hao+Zhang%2C+Mengsi+Lyu%2C+Bo+Huang%2C+Yulong+Ao%2C+and+Yonghua+Lin%2E+2025c%2E+Trimtokenator%2Dlc%3A+Towards+adaptive+visual+token+pruning+for+large+multimodal+models+with+long+contexts%2E+arXiv+preprint+arXiv%3A2512%2E22)
84. Mingyu Zhang, Zixu Li, Zhiwei Chen, Zhiheng Fu, Xiaowei Zhu, Jiajia Nie, Yinwei Wei, and Yupeng Hu. 2026a. Hint: Composed image retrieval with dual-path compositional contextualized network. arXiv preprint arXiv:2603.26341. — [arXiv:2603.26341](https://arxiv.org/abs/2603.26341) [find](https://scholar.google.com/scholar?q=Mingyu+Zhang%2C+Zixu+Li%2C+Zhiwei+Chen%2C+Zhiheng+Fu%2C+Xiaowei+Zhu%2C+Jiajia+Nie%2C+Yinwei+Wei%2C+and+Yupeng+Hu%2E+2026a%2E+Hint%3A+Composed+image+retrieval+with+dual%2Dpath+compositional+contextualized+network%2E+arXiv+pre)
85. Qianchi Zhang, Hainan Zhang, Liang Pang, Hongwei Zheng, and Zhiming Zheng. 2026b. [Stable-rag: Mitigating retrieval-permutation-induced hallucinations in retrieval-augmented generation](https://md.research.elamur.ai/free/en/papers/zhang-2026-stable-rag-mitigating-retrieval-permutation/). arXiv preprint arXiv:2601.02993.
86. Zhenyu Zhang, Ying Sheng, Tianyi Zhou, Tianlong Chen, Lianmin Zheng, Ruisi Cai, Zhao Song, Yuandong Tian, Christopher Ré, Clark W. Barrett, Zhangyang Wang, and Beidi Chen. 2023. H2o: Heavy-hitter oracle for efficient generative inference of large language models. Advances in Neural Information Proce… [find](https://scholar.google.com/scholar?q=Zhenyu+Zhang%2C+Ying+Sheng%2C+Tianyi+Zhou%2C+Tianlong+Chen%2C+Lianmin+Zheng%2C+Ruisi+Cai%2C+Zhao+Song%2C+Yuandong+Tian%2C+Christopher+R%C3%A9%2C+Clark+W%2E+Barrett%2C+Zhangyang+Wang%2C+and+Beidi+Chen%2E+2023%2E+H2o%3A+Heavy%2Dhitter+orac)
87. Yao Zhao, Zhitian Xie, Chen Liang, Chenyi Zhuang, and Jinjie Gu. 2024. Lookahead: An inference acceleration framework for large language model with lossless generation accuracy. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 6344–6355. [find](https://scholar.google.com/scholar?q=Yao+Zhao%2C+Zhitian+Xie%2C+Chen+Liang%2C+Chenyi+Zhuang%2C+and+Jinjie+Gu%2E+2024%2E+Lookahead%3A+An+inference+acceleration+framework+for+large+language+model+with+lossless+generation+accuracy%2E+In+Proceedings+of+the+)
88. Qinkai Zheng, Xiao Xia, Xu Zou, Yuxiao Dong, Shan Wang, Yufei Xue, Lei Shen, Zihan Wang, Andi Wang, Yang Li, Teng Su, Zhilin Yang, and Jie Tang. 2023. [Codegeex: A pre-trained model for code generation with multilingual benchmarking on humaneval-x](https://md.research.elamur.ai/free/en/papers/zheng-2023-codegeex-pre-trained-model-code/). In Proceedings of the 29th ACM SIGKDD Conference on…
89. Wangchunshu Zhou, Canwen Xu, Tao Ge, Julian McAuley, Ke Xu, and Furu Wei. 2020. BERT loses patience: Fast and robust inference with early exit. arXiv preprint arXiv:2006.04152. — [arXiv:2006.04152](https://arxiv.org/abs/2006.04152) [find](https://scholar.google.com/scholar?q=Wangchunshu+Zhou%2C+Canwen+Xu%2C+Tao+Ge%2C+Julian+McAuley%2C+Ke+Xu%2C+and+Furu+Wei%2E+2020%2E+BERT+loses+patience%3A+Fast+and+robust+inference+with+early+exit%2E+arXiv+preprint+arXiv%3A2006%2E04152%2E+%E2%80%94+%5BarXiv%3A2006%2E04152%5D%28ht)

*Updated 31 October 2026*