---
title: "Spotlight Your Instructions: Instruction-following with Dynamic"
description: "The authors propose SpotLight, an inference-time technique. 57 references, cited by 4 papers here."
authors: ["Praveen Venkateswaran", "Danish Contractor"]
year: 2026
venue: "Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)"
source: "https://aclanthology.org/2026.eacl-long.174/"
doi: "10.18653/v1/2026.eacl-long.174"
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/venkateswaran-2026-spotlight-your-instructions-instruction-following/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/venkateswaran-2026-spotlight-your-instructions-instruction-following/)

[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/page/4/) / Spotlight Your Instructions: Instruction-following with Dynamic Attention Steering

# Spotlight Your Instructions: Instruction-following with Dynamic Attention Steering

**Authors:** [Praveen Venkateswaran](https://md.research.elamur.ai/free/en/authors/praveen-venkateswaran/) and [Danish Contractor](https://md.research.elamur.ai/free/en/authors/danish-contractor/) · **Year:** 2026 · **Venue:** [Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)](https://md.research.elamur.ai/free/en/venues/eacl/) ([EACL 2026](https://md.research.elamur.ai/free/en/venues/eacl/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.eacl-long.174/) · [PDF](https://aclanthology.org/2026.eacl-long.174.pdf) · [DOI 10.18653/v1/2026.eacl-long.174](https://doi.org/10.18653/v1/2026.eacl-long.174) · [OpenAlex](https://openalex.org/W7140117179)

**Citations:** 0 citations (OpenAlex, 2026-10-14) · [4](https://md.research.elamur.ai/free/en/papers/venkateswaran-2026-spotlight-your-instructions-instruction-following/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
LLMs frequently overlook explicit instructions, especially when prompts are long or contain many constraints. Users have few ways to stress what matters beyond rewording the prompt. Earlier attention-steering methods such as PASTA add a fixed bias and need costly per-model, per-task profiling to choose heads and layers. Fixed biases can also over-steer and make outputs incoherent.

The authors propose SpotLight, an inference-time technique. For each query position, it measures the share of attention that falls on user-marked prompt spans. If that share is below a chosen target, it adds a logarithmic bias to the span tokens' attention logits, scaled to the size of the deficit. Steering is applied in every layer and head, so no profiling or fine-tuning is required.

SpotLight was evaluated on four task types with seven models (3B–72B). It improved instruction adherence while keeping, and sometimes raising, task quality, at a modest latency cost. This makes per-prompt emphasis practical in real deployments.

## Key points
- **Bias rule:** the added bias equals log(ψ_target / ψ_current) for span tokens and zero otherwise.
  - The bias shrinks toward zero when the model already attends enough, which prevents over-steering.
  - Rankings within span tokens and within non-span tokens are preserved.
  - A proposition bounds the resulting attention share between ψ_target/(1+ψ_target) and ψ_target.
- **Instruction-following benchmarks:**
  - IFEval: 541 prompts, 25 instruction types, 1–3 instructions each. Prompts were restructured with Mixtral-8x22b-Instruct to separate tasks from instructions.
  - ManyIFEval: 1000 instances with 10 instructions each.
  - MT-IFEval: a new multi-turn set of 300 five-turn conversations.
- **Task-capability checks:**
  - HotpotQA: 1000 sampled distractor-setting questions.
  - ComplexBench.
  - Safety-related evaluations: WildJailbreak and CoCoNot.
- **Illustrative example:** with Qwen2.5 7B Instruct and ψ_target = 0.1, SpotLight turned a non-compliant answer into a fully compliant one. An extreme ψ_target produced incoherent text.
- **Implementation and ablations:**
  - Released as a plug-and-play hook for HuggingFace Transformers and timed on a single A100 80GB.
  - Steering attention logits outperformed directly reweighting attention probabilities.
  - A known limitation is that some caching optimizations cannot be used, which increases memory 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.*

## Institutions

- [IBM](https://md.research.elamur.ai/free/en/companies/ibm/) — Company, Armonk, United States · 139 works · 70,691 citations

📚 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 · [👁️ Computer Vision and Pattern Recognition](https://md.research.elamur.ai/free/en/shelves/computer-science/computer-vision-and-pattern-recognition/) · 2,869 works

## Cite

Cite as: Praveen Venkateswaran and Danish Contractor. 2026. Spotlight Your Instructions: Instruction-following with Dynamic Attention Steering. Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers).

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

```bibtex
@inproceedings{venkateswaran2026spotlight,
  title = {{Spotlight Your Instructions: Instruction-following with Dynamic Attention Steering}},
  author = {Praveen Venkateswaran and Danish Contractor},
  booktitle = {Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.eacl-long.174},
  url = {https://doi.org/10.18653/v1/2026.eacl-long.174}
}
```

- [← previous Sparse-RL: Breaking the Memory Wall in LLM Reinforcement Learning via Stable Sparse Rollouts](https://md.research.elamur.ai/free/en/papers/luo-2026-sparse-rl-breaking-memory-wall/)
- [next → Standard-to-Dialect Transfer Trends Differ across Text and Speech: A Case Study on Intent and Topic Classification in German Dialects](https://md.research.elamur.ai/free/en/papers/blaschke-2026-standard-dialect-transfer-trends-differ/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/4/)

## References

57 references, 19 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Ibrahim Abdelaziz, Kinjal Basu, Mayank Agarwal, Sadhana Kumaravel, Matthew Stallone, Rameswar Panda, Yara Rizk, GP Bhargav, Maxwell Crouse, Chulaka Gunasekara, and 1 others. 2024. Granite-function calling model: Introducing function calling abilities via multi-task learning of granular tasks. arXiv… [find](https://scholar.google.com/scholar?q=Ibrahim+Abdelaziz%2C+Kinjal+Basu%2C+Mayank+Agarwal%2C+Sadhana+Kumaravel%2C+Matthew+Stallone%2C+Rameswar+Panda%2C+Yara+Rizk%2C+GP+Bhargav%2C+Maxwell+Crouse%2C+Chulaka+Gunasekara%2C+and+1+others%2E+2024%2E+Granite%2Dfunction+cal)
2. Yuntao Bai, Andy Jones, Kamal Ndousse, Amanda Askell, Anna Chen, Nova DasSarma, Dawn Drain, Stanislav Fort, Deep Ganguli, Tom Henighan, and 1 others. 2022. Training a helpful and harmless assistant with reinforcement learning from human feedback. arXiv preprint arXiv:2204.05862. [find](https://scholar.google.com/scholar?q=Yuntao+Bai%2C+Andy+Jones%2C+Kamal+Ndousse%2C+Amanda+Askell%2C+Anna+Chen%2C+Nova+DasSarma%2C+Dawn+Drain%2C+Stanislav+Fort%2C+Deep+Ganguli%2C+Tom+Henighan%2C+and+1+others%2E+2022%2E+Training+a+helpful+and+harmless+assistant+wi)
3. Amit Ben-Artzy and Roy Schwartz. 2024. Attend first, consolidate later: On the importance of attention in different llm layers. arXiv preprint arXiv:2409.03621. [find](https://scholar.google.com/scholar?q=Amit+Ben%2DArtzy+and+Roy+Schwartz%2E+2024%2E+Attend+first%2C+consolidate+later%3A+On+the+importance+of+attention+in+different+llm+layers%2E+arXiv+preprint+arXiv%3A2409%2E03621%2E)
4. Luca Beurer-Kellner, Marc Fischer, and Martin Vechev. 2024. Guiding llms the right way: Fast, non-invasive constrained generation. arXiv preprint arXiv:2403.06988. [find](https://scholar.google.com/scholar?q=Luca+Beurer%2DKellner%2C+Marc+Fischer%2C+and+Martin+Vechev%2E+2024%2E+Guiding+llms+the+right+way%3A+Fast%2C+non%2Dinvasive+constrained+generation%2E+arXiv+preprint+arXiv%3A2403%2E06988%2E)
5. Faeze Brahman, Sachin Kumar, Vidhisha Balachandran, Pradeep Dasigi, Valentina Pyatkin, Abhilasha Ravichander, Sarah Wiegreffe, Nouha Dziri, Khyathi Chandu, Jack Hessel, and 1 others. 2024. [The art of saying no: Contextual noncompliance in language models](https://md.research.elamur.ai/free/en/papers/brahman-2024-art-saying-no-contextual-noncompliance/). Advances in Neural Information Processing Sy…
6. Sizhe Chen, Julien Piet, Chawin Sitawarin, and David Wagner. 2024. Struq: Defending against prompt injection with structured queries. arXiv preprint arXiv:2402.06363. — [arXiv:2402.06363](https://arxiv.org/abs/2402.06363) [find](https://scholar.google.com/scholar?q=Sizhe+Chen%2C+Julien+Piet%2C+Chawin+Sitawarin%2C+and+David+Wagner%2E+2024%2E+Struq%3A+Defending+against+prompt+injection+with+structured+queries%2E+arXiv+preprint+arXiv%3A2402%2E06363%2E+%E2%80%94+%5BarXiv%3A2402%2E06363%5D%28https%3A%2F%2Farxi)
7. Hyung Won Chung, Le Hou, Shayne Longpre, Barret Zoph, Yi Tay, William Fedus, Yunxuan Li, Xuezhi Wang, Mostafa Dehghani, Siddhartha Brahma, and 1 others. 2024. Scaling instruction-finetuned language models. Journal of Machine Learning Research, 25(70):1–53. [find](https://scholar.google.com/scholar?q=Hyung+Won+Chung%2C+Le+Hou%2C+Shayne+Longpre%2C+Barret+Zoph%2C+Yi+Tay%2C+William+Fedus%2C+Yunxuan+Li%2C+Xuezhi+Wang%2C+Mostafa+Dehghani%2C+Siddhartha+Brahma%2C+and+1+others%2E+2024%2E+Scaling+instruction%2Dfinetuned+language+mo)
8. Björn Deiseroth, Mayukh Deb, Samuel Weinbach, Manuel Brack, Patrick Schramowski, and Kristian Kersting. 2023. Atman: Understanding transformer predictions through memory efficient attention manipulation. Advances in Neural Information Processing Systems, 36:63437–63460. [find](https://scholar.google.com/scholar?q=Bj%C3%B6rn+Deiseroth%2C+Mayukh+Deb%2C+Samuel+Weinbach%2C+Manuel+Brack%2C+Patrick+Schramowski%2C+and+Kristian+Kersting%2E+2023%2E+Atman%3A+Understanding+transformer+predictions+through+memory+efficient+attention+manipulati)
9. Yue Dong, Chandra Bhagavatula, Ximing Lu, Jena D Hwang, Antoine Bosselut, Jackie Chi Kit Cheung, and Yejin Choi. 2021. On-the-fly attention modulation for neural generation. arXiv preprint arXiv:2101.00371. [find](https://scholar.google.com/scholar?q=Yue+Dong%2C+Chandra+Bhagavatula%2C+Ximing+Lu%2C+Jena+D+Hwang%2C+Antoine+Bosselut%2C+Jackie+Chi+Kit+Cheung%2C+and+Yejin+Choi%2E+2021%2E+On%2Dthe%2Dfly+attention+modulation+for+neural+generation%2E+arXiv+preprint+arXiv%3A2101%2E)
10. Konstantin Donhauser, Charles Arnal, Mohammad Pezeshki, Vivien Cabannes, David Lopez-Paz, and Kartik Ahuja. 2025. Unveiling simplicities of attention: Adaptive long-context head identification. arXiv preprint arXiv:2502.09647. — [arXiv:2502.09647](https://arxiv.org/abs/2502.09647) [find](https://scholar.google.com/scholar?q=Konstantin+Donhauser%2C+Charles+Arnal%2C+Mohammad+Pezeshki%2C+Vivien+Cabannes%2C+David+Lopez%2DPaz%2C+and+Kartik+Ahuja%2E+2025%2E+Unveiling+simplicities+of+attention%3A+Adaptive+long%2Dcontext+head+identification%2E+arXiv+)
11. IBM Granite Team. 2024. Granite 3.0 language models. [find](https://scholar.google.com/scholar?q=IBM+Granite+Team%2E+2024%2E+Granite+3%2E0+language+models%2E)
12. Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Alex Vaughan, and 1 others. 2024. [The llama 3 herd of models](https://md.research.elamur.ai/free/en/papers/grattafiori-2024-llama-3-herd-models/). arXiv preprint arXiv:2407.21783.
13. Qin Guo and Tianwei Lin. 2024. Focus on your instruction: Fine-grained and multi-instruction image editing by attention modulation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6986–6996. [find](https://scholar.google.com/scholar?q=Qin+Guo+and+Tianwei+Lin%2E+2024%2E+Focus+on+your+instruction%3A+Fine%2Dgrained+and+multi%2Dinstruction+image+editing+by+attention+modulation%2E+In+Proceedings+of+the+IEEE%2FCVF+Conference+on+Computer+Vision+and+Pat)
14. Keno Harada, Yudai Yamazaki, Masachika Taniguchi, Edison Marrese-Taylor, Takeshi Kojima, Yusuke Iwasawa, and Yutaka Matsuo. 2025. When instructions multiply: Measuring and estimating llm capabilities of multiple instructions following. Preprint, arXiv:2509.21051. [find](https://scholar.google.com/scholar?q=Keno+Harada%2C+Yudai+Yamazaki%2C+Masachika+Taniguchi%2C+Edison+Marrese%2DTaylor%2C+Takeshi+Kojima%2C+Yusuke+Iwasawa%2C+and+Yutaka+Matsuo%2E+2025%2E+When+instructions+multiply%3A+Measuring+and+estimating+llm+capabilities+)
15. Thomas Hartvigsen, Saadia Gabriel, Hamid Palangi, Maarten Sap, Dipankar Ray, and Ece Kamar. 2022. [Toxigen: A large-scale machine-generated dataset for adversarial and implicit hate speech detection](https://md.research.elamur.ai/free/en/papers/hartvigsen-2022-toxigen-large-scale-machine-generated/). arXiv preprint arXiv:2203.09509.
16. Jerry Zhi-Yang He, Sashrika Pandey, Mariah L Schrum, and Anca Dragan. 2024. Context steering: Controllable personalization at inference time. arXiv preprint arXiv:2405.01768. — [arXiv:2405.01768](https://arxiv.org/abs/2405.01768) [find](https://scholar.google.com/scholar?q=Jerry+Zhi%2DYang+He%2C+Sashrika+Pandey%2C+Mariah+L+Schrum%2C+and+Anca+Dragan%2E+2024%2E+Context+steering%3A+Controllable+personalization+at+inference+time%2E+arXiv+preprint+arXiv%3A2405%2E01768%2E+%E2%80%94+%5BarXiv%3A2405%2E01768%5D%28http)
17. Juyeon Heo, Christina Heinze-Deml, Oussama Elachqar, Shirley Ren, Udhay Nallasamy, Andy Miller, Kwan Ho Ryan Chan, and Jaya Narain. 2024. Do llms" know" internally when they follow instructions? arXiv preprint arXiv:2410.14516. — [arXiv:2410.14516](https://arxiv.org/abs/2410.14516) [find](https://scholar.google.com/scholar?q=Juyeon+Heo%2C+Christina+Heinze%2DDeml%2C+Oussama+Elachqar%2C+Shirley+Ren%2C+Udhay+Nallasamy%2C+Andy+Miller%2C+Kwan+Ho+Ryan+Chan%2C+and+Jaya+Narain%2E+2024%2E+Do+llms%22+know%22+internally+when+they+follow+instructions%3F+arXiv)
18. Cheng-Yu Hsieh, Yung-Sung Chuang, Chun-Liang Li, Zifeng Wang, Long Le, Abhishek Kumar, James Glass, Alexander Ratner, Chen-Yu Lee, Ranjay Krishna, and Tomas Pfister. 2024. Found in the middle: Calibrating positional attention bias improves long context utilization. In Findings of the Association for… [find](https://scholar.google.com/scholar?q=Cheng%2DYu+Hsieh%2C+Yung%2DSung+Chuang%2C+Chun%2DLiang+Li%2C+Zifeng+Wang%2C+Long+Le%2C+Abhishek+Kumar%2C+James+Glass%2C+Alexander+Ratner%2C+Chen%2DYu+Lee%2C+Ranjay+Krishna%2C+and+Tomas+Pfister%2E+2024%2E+Found+in+the+middle%3A+Calibra)
19. Lei Huang, Weijiang Yu, Weitao Ma, Weihong Zhong, Zhangyin Feng, Haotian Wang, Qianglong Chen, Weihua Peng, Xiaocheng Feng, Bing Qin, and 1 others. 2025. [A survey on hallucination in large language models: Principles, taxonomy, challenges, and open questions](https://md.research.elamur.ai/free/en/papers/huang-2023-survey-hallucination-large-language-models/). ACM Transactions on Information Systems,…
20. Jiabao Ji, Yoon Kim, James Glass, and Tianxing He. 2022. Controlling the focus of pre-trained language generation models. arXiv preprint arXiv:2203.01146. — [arXiv:2203.01146](https://arxiv.org/abs/2203.01146) [find](https://scholar.google.com/scholar?q=Jiabao+Ji%2C+Yoon+Kim%2C+James+Glass%2C+and+Tianxing+He%2E+2022%2E+Controlling+the+focus+of+pre%2Dtrained+language+generation+models%2E+arXiv+preprint+arXiv%3A2203%2E01146%2E+%E2%80%94+%5BarXiv%3A2203%2E01146%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F22)
21. Albert Q. Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lucile Saulnier, Lélio Renard Lavaud, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas Wang, Timothée Lacroix,… [find](https://scholar.google.com/scholar?q=Albert+Q%2E+Jiang%2C+Alexandre+Sablayrolles%2C+Arthur+Mensch%2C+Chris+Bamford%2C+Devendra+Singh+Chaplot%2C+Diego+de+las+Casas%2C+Florian+Bressand%2C+Gianna+Lengyel%2C+Guillaume+Lample%2C+Lucile+Saulnier%2C+L%C3%A9lio+Renard+Lav)
22. Albert Q Jiang, Alexandre Sablayrolles, Antoine Roux, Arthur Mensch, Blanche Savary, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Emma Bou Hanna, Florian Bressand, and 1 others. 2024a. Mixtral of experts. arXiv preprint arXiv:2401.04088. [find](https://scholar.google.com/scholar?q=Albert+Q+Jiang%2C+Alexandre+Sablayrolles%2C+Antoine+Roux%2C+Arthur+Mensch%2C+Blanche+Savary%2C+Chris+Bamford%2C+Devendra+Singh+Chaplot%2C+Diego+de+las+Casas%2C+Emma+Bou+Hanna%2C+Florian+Bressand%2C+and+1+others%2E+2024a%2E+M)
23. Liwei Jiang, Kavel Rao, Seungju Han, Allyson Ettinger, Faeze Brahman, Sachin Kumar, Niloofar Mireshghallah, Ximing Lu, Maarten Sap, Yejin Choi, and 1 others. 2024b. [Wildteaming at scale: From in-the-wild jailbreaks to (adversarially) safer language models](https://md.research.elamur.ai/free/en/papers/jiang-2024-wildteaming-scale-wild-jailbreaks-adversarially/). Advances in Neural Information Processing S…
24. Wang-Cheng Kang, Jianmo Ni, Nikhil Mehta, Maheswaran Sathiamoorthy, Lichan Hong, Ed Chi, and Derek Zhiyuan Cheng. 2023. Do llms understand user preferences? evaluating llms on user rating prediction. arXiv preprint arXiv:2305.06474. — [arXiv:2305.06474](https://arxiv.org/abs/2305.06474) [find](https://scholar.google.com/scholar?q=Wang%2DCheng+Kang%2C+Jianmo+Ni%2C+Nikhil+Mehta%2C+Maheswaran+Sathiamoorthy%2C+Lichan+Hong%2C+Ed+Chi%2C+and+Derek+Zhiyuan+Cheng%2E+2023%2E+Do+llms+understand+user+preferences%3F+evaluating+llms+on+user+rating+prediction%2E+)
25. Bruce W Lee, Inkit Padhi, Karthikeyan Natesan Ramamurthy, Erik Miehling, Pierre Dognin, Manish Nagireddy, and Amit Dhurandhar. 2024. Programming refusal with conditional activation steering. arXiv preprint arXiv:2409.05907. [find](https://scholar.google.com/scholar?q=Bruce+W+Lee%2C+Inkit+Padhi%2C+Karthikeyan+Natesan+Ramamurthy%2C+Erik+Miehling%2C+Pierre+Dognin%2C+Manish+Nagireddy%2C+and+Amit+Dhurandhar%2E+2024%2E+Programming+refusal+with+conditional+activation+steering%2E+arXiv+pre)
26. Kenneth Li, Tianle Liu, Naomi Bashkansky, David Bau, Fernanda Viégas, Hanspeter Pfister, and Martin Wattenberg. 2024. Measuring and controlling instruction (in)stability in language model dialogs. Preprint, arXiv:2402.10962. — [arXiv:2402.10962](https://arxiv.org/abs/2402.10962) [find](https://scholar.google.com/scholar?q=Kenneth+Li%2C+Tianle+Liu%2C+Naomi+Bashkansky%2C+David+Bau%2C+Fernanda+Vi%C3%A9gas%2C+Hanspeter+Pfister%2C+and+Martin+Wattenberg%2E+2024%2E+Measuring+and+controlling+instruction+%28in%29stability+in+language+model+dialogs%2E+Pre)
27. Chris Yuhao Liu, Liang Zeng, Yuzhen Xiao, Jujie He, Jiacai Liu, Chaojie Wang, Rui Yan, Wei Shen, Fuxiang Zhang, Jiacheng Xu, and 1 others. 2025. [Skywork-reward-v2: Scaling preference data curation via human-ai synergy](https://md.research.elamur.ai/free/en/papers/liu-2025-skywork-reward-v2-scaling/). arXiv preprint arXiv:2507.01352.
28. Michael Xieyang Liu, Frederick Liu, Alexander J Fiannaca, Terry Koo, Lucas Dixon, Michael Terry, and Carrie J Cai. 2024a. " we need structured output": Towards user-centered constraints on large language model output. In Extended Abstracts of the CHI Conference on Human Factors in Computing Systems,… [find](https://scholar.google.com/scholar?q=Michael+Xieyang+Liu%2C+Frederick+Liu%2C+Alexander+J+Fiannaca%2C+Terry+Koo%2C+Lucas+Dixon%2C+Michael+Terry%2C+and+Carrie+J+Cai%2E+2024a%2E+%22+we+need+structured+output%22%3A+Towards+user%2Dcentered+constraints+on+large+langu)
29. Nelson F Liu, Kevin Lin, John Hewitt, Ashwin Paranjape, Michele Bevilacqua, Fabio Petroni, and Percy Liang. 2024b. [Lost in the middle: How language models use long contexts](https://md.research.elamur.ai/free/en/papers/liu-2023-lost-middle-language-models-use/). Transactions of the Association for Computational Linguistics, 12:157–173.
30. Raja Sekhar Reddy Mekala, Yasaman Razeghi, and Sameer Singh. 2024. EchoPrompt: Instructing the model to rephrase queries for improved in-context learning. In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies… [find](https://scholar.google.com/scholar?q=Raja+Sekhar+Reddy+Mekala%2C+Yasaman+Razeghi%2C+and+Sameer+Singh%2E+2024%2E+EchoPrompt%3A+Instructing+the+model+to+rephrase+queries+for+improved+in%2Dcontext+learning%2E+In+Proceedings+of+the+2024+Conference+of+the+)
31. Sean O’Brien and Mike Lewis. 2023. [Contrastive decoding improves reasoning in large language models](https://md.research.elamur.ai/free/en/papers/o-brien-2023-contrastive-decoding-improves-reasoning-large/). arXiv preprint arXiv:2309.09117.
32. Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, and 1 others. 2022. [Training language models to follow instructions with human feedback](https://md.research.elamur.ai/free/en/papers/ouyang-2022-training-language-models-follow-instructions/). Advances in neural information processing systems, 35:27730–27744.
33. Baolin Peng, Chunyuan Li, Pengcheng He, Michel Galley, and Jianfeng Gao. 2023. [Instruction tuning with gpt-4](https://md.research.elamur.ai/free/en/papers/peng-2023-instruction-tuning-gpt-4/). arXiv preprint arXiv:2304.03277.
34. Rui Pu, Chaozhuo Li, Rui Ha, Zejian Chen, Litian Zhang, Zheng Liu, Lirong Qiu, and Xi Zhang. 2024. Feint and attack: Attention-based strategies for jailbreaking and protecting llms. arXiv preprint arXiv:2410.16327. — [arXiv:2410.16327](https://arxiv.org/abs/2410.16327) [find](https://scholar.google.com/scholar?q=Rui+Pu%2C+Chaozhuo+Li%2C+Rui+Ha%2C+Zejian+Chen%2C+Litian+Zhang%2C+Zheng+Liu%2C+Lirong+Qiu%2C+and+Xi+Zhang%2E+2024%2E+Feint+and+attack%3A+Attention%2Dbased+strategies+for+jailbreaking+and+protecting+llms%2E+arXiv+preprint+arX)
35. Yiwei Qin, Kaiqiang Song, Yebowen Hu, Wenlin Yao, Sangwoo Cho, Xiaoyang Wang, Xuansheng Wu, Fei Liu, Pengfei Liu, and Dong Yu. 2024. [Infobench: Evaluating instruction following ability in large language models](https://md.research.elamur.ai/free/en/papers/qin-2024-infobench-evaluating-instruction-following-ability/). In Findings of the Association for Computational Linguistics ACL 2024, pages 13025– 13048…
36. Jérémy Scheurer, Jon Ander Campos, Tomasz Korbak, Jun Shern Chan, Angelica Chen, Kyunghyun Cho, and Ethan Perez. 2023. Training language models with language feedback at scale. arXiv preprint arXiv:2303.16755. [find](https://scholar.google.com/scholar?q=J%C3%A9r%C3%A9my+Scheurer%2C+Jon+Ander+Campos%2C+Tomasz+Korbak%2C+Jun+Shern+Chan%2C+Angelica+Chen%2C+Kyunghyun+Cho%2C+and+Ethan+Perez%2E+2023%2E+Training+language+models+with+language+feedback+at+scale%2E+arXiv+preprint+arXiv%3A23)
37. Xinyue Shen, Zeyuan Chen, Michael Backes, Yun Shen, and Yang Zhang. 2024. [" do anything now": Characterizing and evaluating in-the-wild jailbreak prompts on large language models](https://md.research.elamur.ai/free/en/papers/shen-2023-anything-now-characterizing-evaluating-wild/). In Proceedings of the 2024 on ACM SIGSAC Conference on Computer and Communications Security, pages 1671–1685.
38. Chandan Singh, Aliyah R Hsu, Richard Antonello, Shailee Jain, Alexander G Huth, Bin Yu, and Jianfeng Gao. 2023. Explaining black box text modules in natural language with language models. arXiv preprint arXiv:2305.09863. — [arXiv:2305.09863](https://arxiv.org/abs/2305.09863) [find](https://scholar.google.com/scholar?q=Chandan+Singh%2C+Aliyah+R+Hsu%2C+Richard+Antonello%2C+Shailee+Jain%2C+Alexander+G+Huth%2C+Bin+Yu%2C+and+Jianfeng+Gao%2E+2023%2E+Explaining+black+box+text+modules+in+natural+language+with+language+models%2E+arXiv+prepri)
39. Alessandro Stolfo, Vidhisha Balachandran, Safoora Yousefi, Eric Horvitz, and Besmira Nushi. 2024. Improving instruction-following in language models through activation steering. arXiv preprint arXiv:2410.12877. — [arXiv:2410.12877](https://arxiv.org/abs/2410.12877) [find](https://scholar.google.com/scholar?q=Alessandro+Stolfo%2C+Vidhisha+Balachandran%2C+Safoora+Yousefi%2C+Eric+Horvitz%2C+and+Besmira+Nushi%2E+2024%2E+Improving+instruction%2Dfollowing+in+language+models+through+activation+steering%2E+arXiv+preprint+arXiv%3A2)
40. Dimitri Von Rütte, Sotiris Anagnostidis, Gregor Bachmann, and Thomas Hofmann. 2024. A language model’s guide through latent space. arXiv preprint arXiv:2402.14433. — [arXiv:2402.14433](https://arxiv.org/abs/2402.14433) [find](https://scholar.google.com/scholar?q=Dimitri+Von+R%C3%BCtte%2C+Sotiris+Anagnostidis%2C+Gregor+Bachmann%2C+and+Thomas+Hofmann%2E+2024%2E+A+language+model%E2%80%99s+guide+through+latent+space%2E+arXiv+preprint+arXiv%3A2402%2E14433%2E+%E2%80%94+%5BarXiv%3A2402%2E14433%5D%28https%3A%2F%2Farxiv%2Eo)
41. Eric Wallace, Kai Xiao, Reimar Leike, Lilian Weng, Johannes Heidecke, and Alex Beutel. 2024. The instruction hierarchy: Training llms to prioritize privileged instructions. arXiv preprint arXiv:2404.13208. [find](https://scholar.google.com/scholar?q=Eric+Wallace%2C+Kai+Xiao%2C+Reimar+Leike%2C+Lilian+Weng%2C+Johannes+Heidecke%2C+and+Alex+Beutel%2E+2024%2E+The+instruction+hierarchy%3A+Training+llms+to+prioritize+privileged+instructions%2E+arXiv+preprint+arXiv%3A2404%2E1)
42. Alexander Wei, Nika Haghtalab, and Jacob Steinhardt. 2023. [Jailbroken: How does llm safety training fail?](https://md.research.elamur.ai/free/en/papers/wei-2023-jailbroken-llm-safety-training-fail/) Advances in Neural Information Processing Systems, 36:80079–80110.
43. Jason Wei, Maarten Bosma, Vincent Y Zhao, Kelvin Guu, Adams Wei Yu, Brian Lester, Nan Du, Andrew M Dai, and Quoc V Le. 2021. [Finetuned language models are zero-shot learners](https://md.research.elamur.ai/free/en/papers/wei-2021-finetuned-language-models-zero-shot/). arXiv preprint arXiv:2109.01652.
44. Bosi Wen, Pei Ke, Xiaotao Gu, Lindong Wu, Hao Huang, Jinfeng Zhou, Wenchuang Li, Binxin Hu, Wendy Gao, Jiaxing Xu, and 1 others. 2024. Benchmarking complex instruction-following with multiple constraints composition. Advances in Neural Information Processing Systems, 37:137610–137645. [find](https://scholar.google.com/scholar?q=Bosi+Wen%2C+Pei+Ke%2C+Xiaotao+Gu%2C+Lindong+Wu%2C+Hao+Huang%2C+Jinfeng+Zhou%2C+Wenchuang+Li%2C+Binxin+Hu%2C+Wendy+Gao%2C+Jiaxing+Xu%2C+and+1+others%2E+2024%2E+Benchmarking+complex+instruction%2Dfollowing+with+multiple+constrai)
45. Brandon T Willard and Rémi Louf. 2023. Efficient guided generation for large language models. arXiv preprint arXiv:2307.09702. [find](https://scholar.google.com/scholar?q=Brandon+T+Willard+and+R%C3%A9mi+Louf%2E+2023%2E+Efficient+guided+generation+for+large+language+models%2E+arXiv+preprint+arXiv%3A2307%2E09702%2E)
46. Jianhao Yan, Yun Luo, and Yue Zhang. 2024. Refutebench: Evaluating refuting instruction-following for large language models. arXiv preprint arXiv:2402.13463. [find](https://scholar.google.com/scholar?q=Jianhao+Yan%2C+Yun+Luo%2C+and+Yue+Zhang%2E+2024%2E+Refutebench%3A+Evaluating+refuting+instruction%2Dfollowing+for+large+language+models%2E+arXiv+preprint+arXiv%3A2402%2E13463%2E)
47. An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, and 1 others. 2024. Qwen2. 5 technical report. arXiv preprint arXiv:2412.15115. [find](https://scholar.google.com/scholar?q=An+Yang%2C+Baosong+Yang%2C+Beichen+Zhang%2C+Binyuan+Hui%2C+Bo+Zheng%2C+Bowen+Yu%2C+Chengyuan+Li%2C+Dayiheng+Liu%2C+Fei+Huang%2C+Haoran+Wei%2C+and+1+others%2E+2024%2E+Qwen2%2E+5+technical+report%2E+arXiv+preprint+arXiv%3A2412%2E15115)
48. Zhilin Yang, Peng Qi, Saizheng Zhang, Yoshua Bengio, William W Cohen, Ruslan Salakhutdinov, and Christopher D Manning. 2018. [Hotpotqa: A dataset for diverse, explainable multi-hop question answering](https://md.research.elamur.ai/free/en/papers/yang-2018-hotpotqa-dataset-diverse-explainable-multi/). arXiv preprint arXiv:1809.09600.
49. Jiao Sun, Yufei Tian, Wangchunshu Zhou, Nan Xu, Qian Hu, Rahul Gupta, John Frederick Wieting, Nanyun Peng, and Xuezhe Ma. 2023. Evaluating large language models on controlled generation tasks. arXiv preprint arXiv:2310.14542. [find](https://scholar.google.com/scholar?q=Jiao+Sun%2C+Yufei+Tian%2C+Wangchunshu+Zhou%2C+Nan+Xu%2C+Qian+Hu%2C+Rahul+Gupta%2C+John+Frederick+Wieting%2C+Nanyun+Peng%2C+and+Xuezhe+Ma%2E+2023%2E+Evaluating+large+language+models+on+controlled+generation+tasks%2E+arXiv+p)
50. Zhiyuan Zeng, Jiatong Yu, Tianyu Gao, Yu Meng, Tanya Goyal, and Danqi Chen. Evaluating large language models at evaluating instruction following. In NeurIPS 2023 Workshop on Instruction Tuning and Instruction Following. [find](https://scholar.google.com/scholar?q=Zhiyuan+Zeng%2C+Jiatong+Yu%2C+Tianyu+Gao%2C+Yu+Meng%2C+Tanya+Goyal%2C+and+Danqi+Chen%2E+Evaluating+large+language+models+at+evaluating+instruction+following%2E+In+NeurIPS+2023+Workshop+on+Instruction+Tuning+and+Ins)
51. Teun van der Weij, Massimo Poesio, and Nandi Schoots. 2024. Extending activation steering to broad skills and multiple behaviours. arXiv preprint arXiv:2403.05767. — [arXiv:2403.05767](https://arxiv.org/abs/2403.05767) [find](https://scholar.google.com/scholar?q=Teun+van+der+Weij%2C+Massimo+Poesio%2C+and+Nandi+Schoots%2E+2024%2E+Extending+activation+steering+to+broad+skills+and+multiple+behaviours%2E+arXiv+preprint+arXiv%3A2403%2E05767%2E+%E2%80%94+%5BarXiv%3A2403%2E05767%5D%28https%3A%2F%2Farxiv%2Eo)
52. Qingru Zhang, Chandan Singh, Liyuan Liu, Xiaodong Liu, Bin Yu, Jianfeng Gao, and Tuo Zhao. 2023a. Tell your model where to attend: Post-hoc attention steering for llms. arXiv preprint arXiv:2311.02262. — [arXiv:2311.02262](https://arxiv.org/abs/2311.02262) [find](https://scholar.google.com/scholar?q=Qingru+Zhang%2C+Chandan+Singh%2C+Liyuan+Liu%2C+Xiaodong+Liu%2C+Bin+Yu%2C+Jianfeng+Gao%2C+and+Tuo+Zhao%2E+2023a%2E+Tell+your+model+where+to+attend%3A+Post%2Dhoc+attention+steering+for+llms%2E+arXiv+preprint+arXiv%3A2311%2E02262)
53. Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. [Attention is all you need](https://md.research.elamur.ai/free/en/papers/vaswani-2017-attention-all-need/). Advances in neural information processing systems, 30.
54. Shengyu Zhang, Linfeng Dong, Xiaoya Li, Sen Zhang, Xiaofei Sun, Shuhe Wang, Jiwei Li, Runyi Hu, Tianwei Zhang, Fei Wu, and 1 others. 2023b. [Instruction tuning for large language models: A survey](https://md.research.elamur.ai/free/en/papers/zhang-2023-instruction-tuning-large-language-models/). arXiv preprint arXiv:2308.10792.
55. Mingqian Zheng, Jiaxin Pei, Lajanugen Logeswaran, Moontae Lee, and David Jurgens. 2024. When” a helpful assistant” is not really helpful: Personas in system prompts do not improve performances of large language models. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 1… [find](https://scholar.google.com/scholar?q=Mingqian+Zheng%2C+Jiaxin+Pei%2C+Lajanugen+Logeswaran%2C+Moontae+Lee%2C+and+David+Jurgens%2E+2024%2E+When%E2%80%9D+a+helpful+assistant%E2%80%9D+is+not+really+helpful%3A+Personas+in+system+prompts+do+not+improve+performances+of+larg)
56. Jeffrey Zhou, Tianjian Lu, Swaroop Mishra, Siddhartha Brahma, Sujoy Basu, Yi Luan, Denny Zhou, and Le Hou. 2023. [Instruction-following evaluation for large language models](https://md.research.elamur.ai/free/en/papers/zhou-2023-instruction-following-evaluation-large-language/). arXiv preprint arXiv:2311.07911.
57. Andy Zou, Long Phan, Sarah Chen, James Campbell, Phillip Guo, Richard Ren, Alexander Pan, Xuwang Yin, Mantas Mazeika, Ann-Kathrin Dombrowski, and 1 others. 2023. [Representation engineering: A top-down approach to ai transparency](https://md.research.elamur.ai/free/en/papers/zou-2023-representation-engineering-top-down-approach/). arXiv preprint arXiv:2310.01405.

*Updated 1 November 2026*