Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Unified Parameter-Efficient Unlearning for LLMs

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Pith's one-line read LLMEraser shows that instance removal, query modification, and response correction for LoRA-finetuned LLMs can each be reduced to a single influence-function-induced adapter edit that tracks retraining within a few percent of performance.

desk verdict A useful influence-function framework for editing LoRA adapters, but the unlearning claim needs direct forgetting tests before it's supported. read the letter →

arxiv 2412.00383 v2 pith:R4ZACYV4 submitted 2024-11-30 cs.AI cs.LG

classification cs.AIcs.LG
keywords machineunlearninglargelanguagemodelsinfluencefunctionsparameter-efficientfine-tuningLoRAHessian-vectorproductinstance-wisedataprivacy
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

This paper proposes LLMEraser, a framework for unlearning individual training instances from a large language model that was fine-tuned with LoRA adapters. It argues that three kinds of unlearning—deleting an instance, editing its query, or correcting its response—can all be expressed as small parameter adjustments to the adapter, and that influence functions can compute those adjustments directly. On benchmark recommendation and multimodal tasks, the edited adapter performs within 0.0038 AUC of a fully retrained model for instance removal, while running about 31 times faster. If correct, this gives a practical way to honor data-deletion requests without storing and retraining on data shards.

What carries the argument

The mechanism is the influence-function parameter-change formula, Equation 8: $\Delta\Theta_\delta \approx \frac{1}{n}\big(\nabla^2_\Theta R(Z;\hat\Theta)\big)^{-1}\big(G(x,y) - G(x+\delta_x, y+\delta_y)\big)$, which converts an unlearning request into a difference of loss gradients times the inverse Hessian of the empirical risk on the training set. The paper's algorithmic contribution is to avoid forming the inverse Hessian: it solves the equivalent convex quadratic program $F(\Delta)=\frac12 \Delta^\top \nabla^2_\Theta R(Z;\hat\Theta)\Delta - \langle b,\Delta\rangle$ using SGD/Adam with Hessian-vector products, so each iteration costs $O(p)$ instead of $O(p^2)$ or $O(p^3)$.

What would settle it

Train a LoRA adapter on a standard benchmark, stop it before convergence, delete 5% of training instances, and apply LLMEraser; if the edited adapter's AUC on the held-out set differs from full retraining by substantially more than the reported 0.0038 gap (for instance, 0.02 or more), the near-retrain claim fails for non-converged adapters. A direct numerical-rank check of the LoRA Hessian at the adapter weights would also reveal whether the invertibility condition behind Equation 8 actually holds.

Watch

Extended reading notes

Core claim

The paper's central claim is that instance-wise unlearning in LoRA-tuned LLMs reduces to one operation: computing the parameter change $\Delta\Theta$ a retrained model would undergo, and adding it to the existing adapter. For instance removal, the change is $\Delta\Theta_{\mathrm{IR}} \approx \frac{1}{n}\big(\nabla^2_\Theta R(Z;\hat\Theta)\big)^{-1}\sum_{(x,y)\in S_{\mathrm{IR}}} G(x,y)$; query modification and response correction use the same inverse-Hessian-times-gradient form with the gradient of the edited instance subtracted. The paper then converts the expensive inverse-Hessian-vector product into the solution of a convex finite-sum quadratic program, solved with mini-batch Adam and Hessian-vector products, cutting per-step cost from $O(p^2)$ to $O(p)$. On LLM4Rec and MLLM benchmarks, the resulting adapter lands within 0.0038 AUC of full retraining for instance removal, within 0.024 and 0.048 average accuracy on the two response-correction benchmarks, and within 0.0079 HitRatio@1 for query modification, while taking about $1.4\times 10^3$ seconds on a query-modification task where retraining takes $5.4\times 10^4$ seconds.

Load-bearing premise

The method assumes the fine-tuned adapter sits exactly at the optimum of the training loss, with an invertible Hessian, so that a first-order Taylor expansion accurately predicts the retrained weights.

Editorial extensions

If this is right

  • Instance removal, query modification, and response correction can all be handled by one framework without retraining or altering the model architecture.
  • Adapter edits cost $O(p)$ time and space per step, scaling linearly with the number of LoRA parameters.
  • On the reported benchmarks, the unlearned model stays within 0.0038 AUC (instance removal), 0.024 and 0.048 average accuracy (response correction on two MLLM benchmarks), and 0.0079 HitRatio@1 (query modification) of full retraining.
  • Unlearning is roughly 31 times faster than retraining on the query-modification task ($1.4\times 10^3$ seconds versus $5.4\times 10^4$ seconds on a single A100 GPU).

Reading between the lines

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

  • If the influence-function estimate stays accurate for adapters trained to convergence, the framework could extend to other PEFT methods such as prefix tuning or adapters, since the derivation only requires a differentiable loss with respect to trainable parameters.
  • The assumption that the old adapter is at the empirical-risk minimizer suggests the method may degrade for early-stopped or under-converged checkpoints; a testable extension would be to measure the AUC gap as a function of training epochs or LoRA rank.
  • Because the formulas only need gradients at the removed and replacement instances, the framework points toward a lightweight audit trail: a provider could log per-instance gradients during fine-tuning and later answer ‘what would the model look like without this record?’ without keeping the full training data.
  • The unified IR/QM/RC taxonomy could be applied to safety scenarios, such as removing a harmful output while correcting the same query toward a safe response, combining two edit types in one parameter change.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The manuscript introduces LLMEraser, an influence-function-based framework for instance-wise unlearning in parameter-efficient fine-tuned large language models (PEFT/LLMs). The approach categorizes unlearning requests into Instance Removal, Query Modification, and Response Correction, and computes adapter parameter updates by solving a finite-sum quadratic program that approximates the inverse Hessian-vector product. The authors evaluate LLMEraser on LLM-based recommendation (TallRec, LLaRA) and multimodal LLM (LLaVA 1.5) tasks, reporting utility metrics (AUC, HitRatio@1, accuracy) and runtime against Retrain, SISA, RecEraser, Gradient Ascent, and E2URec. The reported results show near-retrain utility and 31.25x speedup versus retraining.

Significance. The paper's taxonomy of instance-wise unlearning tasks and its unified influence-function treatment of removal and correction are valuable conceptual contributions. If the forgetting behavior were verified, LLMEraser would be a significant practical tool for privacy-preserving editing of deployed PEFT adapters. The method is mathematically well-grounded in standard influence-function machinery, and the reformulation of the Hessian-vector-product computation as a mini-batch convex optimization problem is a plausible scalability improvement. The authors release code, which aids reproducibility. However, the experimental validation as presented only demonstrates utility preservation and does not directly show that target instances have been forgotten, which weakens the central claim as it stands.

major comments (3)
  1. [§4.2.1, Table 2; §4.2.2, Tables 3–5; §4.3, Figure 3] The experiments measure only utility (AUC, HitRatio@1, accuracy) on retained or corrected test data; there is no direct evaluation of whether the influence of the target instances has been removed. For instance removal, a no-op 'Original' model would also attain near-Retrain utility on the test set, so the small gap between LLMEraser and Retrain (0.0038 AUC in Table 2) does not demonstrate forgetting. The same holds for QM and RC: the reported improvements over the Corrupted baseline show that test utility is restored, but not that the model no longer relies on the specific noisy or outdated tokens. Please add forgetting-specific metrics—e.g., loss or accuracy on the target instances, membership inference, or extraction tests—and report them for each task type. Without such evidence, the central claim of the paper is not supported.
  2. [§3.3, Eq. (12); Appendix D.1] The derivation of the parameter change assumes the old adapter \hat{\Theta} is the exact minimizer of the empirical risk (Eq. 3), so that \nabla R(Z;\hat{\Theta})=0, and that the Hessian \nabla^2 R(Z;\hat{\Theta}) is invertible. These assumptions are unlikely to hold for LoRA adapters trained with early-stopped SGD: the final checkpoint is not a stationary point, and the Hessian of an over-parameterized model is typically singular. The error analysis in Appendix D.1 only bounds the Taylor-expansion residual under these assumptions; it does not quantify the error introduced when the assumptions are violated. Please provide empirical checks (e.g., the gradient norm at the checkpoint, an estimate of the Hessian's smallest eigenvalue, or a small-scale experiment where the Hessian can be computed exactly) or provide a modified error bound that relaxes the assumptions.
  3. [§4.4, Table 6; Appendix C] The runtime comparison reports a single timing per method, without standard deviation or number of runs. The claimed 31.25x speedup is a point estimate, and the runtime of LLMEraser depends on the chosen optimizer hyperparameters and the convergence criterion for solving Eq. (12). Please report the mean and standard deviation of wall-clock time over several runs, state the exact settings (optimizer, learning rate, number of iterations, LoRA rank) used for the timing, and clarify whether the reported time includes data loading, gradient computation, and the optimization loop.
minor comments (5)
  1. [Appendix B, Algorithm 1, lines 6–8] The branches for RC and IM use the wrong sets: line 6 uses S_IM instead of S_RC, and line 7 uses S_RC instead of the query-modification set; please correct these labels.
  2. [§3.2, Eq. (13)] The notation S_IM (or SIM) appears in Eq. (13), but the text and Section 3.2 define the set S_QM for query modification; please align the notation.
  3. [§4.2.1] The phrase 'direct estimation of the parameter changes between the retrained model and the original model' overstates what influence functions compute; the method provides a first-order local approximation to the parameter change, not the actual difference between two separately trained models. Please rephrase.
  4. [Tables 2–5 and Figure 3] All metrics are reported as point estimates without error bars; reporting means and standard deviations over multiple random seeds (and data splits) would make the comparisons statistically meaningful.
  5. [§5, Limitations] The Limitations section acknowledges the need for training set access and gradient information, but it does not mention the absence of direct forgetting verification in the experiments; please add a sentence noting that target-instance loss or membership inference evaluations are needed.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the parameter-change estimates are derived from the training loss via influence functions, and the comparison to Retrain is an external benchmark, not an input to the method.

full rationale

The derivation chain in Eqs. 8-11 computes adapter parameter changes from the empirical risk (Eq. 3), the old adapter, and the target or corrected instances supplied by the unlearning request; none of the comparison values (Retrain AUC, HitRatio@1, accuracy) enter the computation of the parameter change delta. Equations 12-15 rewrite the inverse-Hessian-vector product as an equivalent finite-sum quadratic program, which is an algebraic reformulation rather than a fit to any benchmark. Appendix D bounds the error of the Taylor and optimization approximations relative to the true parameter change, not relative to the reported evaluation metrics. The only self-citation, Ding et al. (2025), is invoked as "inspired by" for the finite-sum reformulation, and the paper supplies its own derivation and convergence analysis, so the citation is not load-bearing. The absence of a direct forgetting metric (e.g., membership inference on removed instances) is a verification gap rather than circularity: even a degenerate no-op method would be distinguishable here, since LLMEraser's IR result (AUC 0.6319) moves toward Retrain (0.6357) and away from Original (0.6400), and the QM/RC results likewise move from the corrupted baseline toward the retrained model. No step in the paper reduces to its own inputs by construction, and the Limitations section honestly states the assumptions (gradient access, Taylor-expansion error) that are correctness concerns, not circular steps.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The method introduces no new entities or forces. It relies on standard influence-function mathematics plus three domain assumptions: optimality of the old adapter, invertible Hessian, and availability of the training data. The main hand-chosen values are the optimizer settings for solving the quadratic program, which are not fully specified.

free parameters (3)
  • Adam learning rate for solving Eq. 12 = not reported precisely
    The optimization of the quadratic program uses Adam with a learning rate delta_lr; the value is not given in the main text, and the quality of the estimated parameter change depends on it.
  • Number of optimization iterations = not reported
    The loop in Algorithm 1 runs until convergence; no tolerance or iteration cap is specified, affecting accuracy and runtime.
  • LoRA rank = 8, 16, 32
    The memory and quality results depend on the rank of the adapter; it is an architecture choice rather than a fitted parameter.
assumptions (4)
  • standard math Taylor expansion of the perturbed risk around the old optimum (Koh and Liang 2017)
    Used to derive Equations 6-8 in Section 3.2.
  • domain assumption The old adapter is the exact empirical risk minimizer of Eq. 3
    Influence-function derivation assumes optimality; SGD-trained adapters are not at the exact optimum.
  • domain assumption The Hessian H_hat is positive definite and invertible
    Assumed in Appendix D.1 to justify the Taylor expansion error bound; likely violated for over-parameterized LoRA.
  • domain assumption The full training set Z is available at unlearning time
    Stated in Section 5 Limitations; the method needs the training data and target gradients.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Unified Parameter-Efficient Unlearning for LLMs." pith.science (2026). https://pith.science/paper/R4ZACYV4

@misc{pith2026241200383,
  author       = {Pith},
  title        = {Pith review of: Unified Parameter-Efficient Unlearning for LLMs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/R4ZACYV4}},
  note         = {Machine review of arXiv:2412.00383}
}
read the original abstract

The advent of Large Language Models (LLMs) has revolutionized natural language processing, enabling advanced understanding and reasoning capabilities across a variety of tasks. Fine-tuning these models for specific domains, particularly through Parameter-Efficient Fine-Tuning (PEFT) strategies like LoRA, has become a prevalent practice due to its efficiency. However, this raises significant privacy and security concerns, as models may inadvertently retain and disseminate sensitive or undesirable information. To address these issues, we introduce a novel instance-wise unlearning framework, LLMEraser, which systematically categorizes unlearning tasks and applies precise parameter adjustments using influence functions. Unlike traditional unlearning techniques that are often limited in scope and require extensive retraining, LLMEraser is designed to handle a broad spectrum of unlearning tasks without compromising model performance. Extensive experiments on benchmark datasets demonstrate that LLMEraser excels in efficiently managing various unlearning scenarios while maintaining the overall integrity and efficacy of the models.

Figures

Figures reproduced from arXiv: 2412.00383 by the authors.

Figure 1
Figure 1. 1a: A brief description of the different types of LLM unlearning tasks. 1b: The framework of exact LLM unlearning method, approximate unlearning method. tion (LLM4Rec) as well as MLLM relation mining tasks to validate the effectiveness of LLMEraser. Our extensive evaluations across these diverse scenarios demonstrate that LLMEraser consistently outperforms the state-of-the-art unlearning methods. 2 PRELIMINARY This … view at source ↗
Figure 2
Figure 2. The framework of LLMEraser. The old adapter is obtained through PEFT on domain [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. 3a: Experimental results of the instance removal task using TallRec as the LLM4Rec model on the BookCrossing dataset, where 5% and 10% of the training data were randomly deleted. 3b: Experimental results of the query modification task using LLaRA as the LLM4Rec model on the MovieLens dataset, where interactions were randomly removed from 5% and 10% of users. parameter changes. Additional experimental results and rel… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Instance Removal Case Study & Query Modification Case Study. [PITH_FULL_IMAGE:figures/full_fig_p023_4.png]
Figure 5
Figure 5. Figure 5: Response Correction Case Study. 24 [PITH_FULL_IMAGE:figures/full_fig_p024_5.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. MPSelectTune: Prompt-type Selection for Fine-tuning improves Concept Unlearning in LLMs

    cs.LG 2026-07 conditional novelty 6.5 of 10

    After multi-prompt multi-task fine-tuning, further training on the worst concept-predicting prompt type yields stronger concept unlearning and higher main-task accuracy than uniform multi-prompt or recent baselines.

Reference graph

Works this paper leans on

105 extracted references · 43 canonical work pages · cited by 1 Pith paper

  1. [1]

    Second order stochastic optimization in linear time

    Naman Agarwal, Brian Bullins, and Elad Hazan. Second order stochastic optimization in linear time. CoRR, abs/1602.03943, 2016

  2. [2]

    Glassman

    Ian Arawjo, Chelse Swoopes, Priyan Vaithilingam, Martin Wattenberg, and Elena L. Glassman. Chainforge: A visual toolkit for prompt engineering and LLM hypothesis testing. In CHI , pp.\ 304:1--304:18. ACM , 2024

  3. [3]

    Tallrec: An effective and efficient tuning framework to align large language model with recommendation

    Keqin Bao, Jizhi Zhang, Yang Zhang, Wenjie Wang, Fuli Feng, and Xiangnan He. Tallrec: An effective and efficient tuning framework to align large language model with recommendation. In RecSys, pp.\ 1007--1014. ACM , 2023

  4. [4]

    Bartlett and Shahar Mendelson

    Peter L. Bartlett and Shahar Mendelson. Rademacher and gaussian complexities: Risk bounds and structural results. J. Mach. Learn. Res., 3: 0 463--482, 2002

  5. [5]

    Influence functions in deep learning are fragile

    Samyadeep Basu, Phillip Pope, and Soheil Feizi. Influence functions in deep learning are fragile. In ICLR . OpenReview.net, 2021

  6. [6]

    A fast iterative shrinkage-thresholding algorithm for linear inverse problems

    Amir Beck and Marc Teboulle. A fast iterative shrinkage-thresholding algorithm for linear inverse problems. SIAM J. Imaging Sci. , 2 0 (1): 0 183--202, 2009

  7. [7]

    Nonlinear programming

    Dimitri P Bertsekas. Nonlinear programming. Journal of the Operational Research Society, 48 0 (3): 0 334--334, 1997

  8. [8]

    Digital forgetting in large language models: A survey of unlearning methods

    Alberto Blanco - Justicia, Najeeb Jebreel, Benet Manzanares - Salor, David S \' a nchez, Josep Domingo - Ferrer, Guillem Collell, and Kuan Eeik Tan. Digital forgetting in large language models: A survey of unlearning methods. CoRR, abs/2404.02062, 2024

Show all 105 references
  1. [9]

    Choquette - Choo, Hengrui Jia, Adelin Travers, Baiwu Zhang, David Lie, and Nicolas Papernot

    Lucas Bourtoule, Varun Chandrasekaran, Christopher A. Choquette - Choo, Hengrui Jia, Adelin Travers, Baiwu Zhang, David Lie, and Nicolas Papernot. Machine unlearning. In SP , pp.\ 141--159. IEEE , 2021

  2. [10]

    Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert - Voss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel M. Ziegler, Jef...

  3. [11]

    Robin D. Burke. Hybrid recommender systems: Survey and experiments. User Model. User Adapt. Interact., 12 0 (4): 0 331--370, 2002

  4. [12]

    Prompting change: exploring prompt engineering in large language model ai and its potential to transform education

    William Cain. Prompting change: exploring prompt engineering in large language model ai and its potential to transform education. TechTrends, 68 0 (1): 0 47--57, 2024

  5. [13]

    Proceedings of the 2nd International Workshop on Information Heterogeneity and Fusion in Recommender Systems, HetRec '11, Chicago, Illinois, USA, October 27, 2011, 2011

    Iv \' a n Cantador, Peter Brusilovsky, and Tsvi Kuflik (eds.). Proceedings of the 2nd International Workshop on Information Heterogeneity and Fusion in Recommender Systems, HetRec '11, Chicago, Illinois, USA, October 27, 2011, 2011. ACM

  6. [14]

    Learning to unlearn: Instance-wise unlearning for pre-trained classifiers

    Sungmin Cha, Sungjun Cho, Dasol Hwang, Honglak Lee, Taesup Moon, and Moontae Lee. Learning to unlearn: Instance-wise unlearning for pre-trained classifiers. In AAAI , pp.\ 11186--11194. AAAI Press, 2024

  7. [15]

    Recommendation unlearning

    Chong Chen, Fei Sun, Min Zhang, and Bolin Ding. Recommendation unlearning. In WWW , pp.\ 2768--2777. ACM , 2022

  8. [16]

    Unlearn what you want to forget: Efficient unlearning for llms

    Jiaao Chen and Diyi Yang. Unlearn what you want to forget: Efficient unlearning for llms. In EMNLP , pp.\ 12041--12052. Association for Computational Linguistics, 2023

  9. [17]

    On softmax direct preference optimization for recommendation

    Yuxin Chen, Junfei Tan, An Zhang, Zhengyi Yang, Leheng Sheng, Enzhi Zhang, Xiang Wang, and Tat - Seng Chua. On softmax direct preference optimization for recommendation. In NeurIPS, 2024

  10. [18]

    Exploring the potential of large language models (llms)in learning on graphs

    Zhikai Chen, Haitao Mao, Hang Li, Wei Jin, Hongzhi Wen, Xiaochi Wei, Shuaiqiang Wang, Dawei Yin, Wenqi Fan, Hui Liu, and Jiliang Tang. Exploring the potential of large language models (llms)in learning on graphs. SIGKDD Explor. , 25 0 (2): 0 42--61, 2023

  11. [19]

    Aakanksha Chowdhery, Sharan Narang, Jacob Devlin, Maarten Bosma, Gaurav Mishra, Adam Roberts, Paul Barham, Hyung Won Chung, Charles Sutton, Sebastian Gehrmann, Parker Schuh, Kensen Shi, Sasha Tsvyashchenko, Joshua Maynez, Abhishek Rao, Parker Barnes, Yi Tay, Noam Shazeer, Vino...

  12. [20]

    Personalized pedagogy through a llm-based recommender system

    Nasrin Dehbozorgi, Mourya Teja Kunuku, and Seyedamin Pouriyeh. Personalized pedagogy through a llm-based recommender system. In AIED Companion (2) , volume 2151 of Communications in Computer and Information Science, pp.\ 63--70. Springer, 2024

  13. [21]

    Qlora: Efficient finetuning of quantized llms

    Tim Dettmers, Artidoro Pagnoni, Ari Holtzman, and Luke Zettlemoyer. Qlora: Efficient finetuning of quantized llms. In NeurIPS, 2023

  14. [22]

    BERT: pre-training of deep bidirectional transformers for language understanding

    Jacob Devlin, Ming - Wei Chang, Kenton Lee, and Kristina Toutanova. BERT: pre-training of deep bidirectional transformers for language understanding. In NAACL-HLT (1) , pp.\ 4171--4186. Association for Computational Linguistics, 2019

  15. [23]

    Addressing delayed feedback in conversion rate prediction via influence functions

    Chenlu Ding, Jiancan Wu, Yancheng Yuan, Junfeng Fang, Cunchun Li, Xiang Wang, and Xiangnan He. Addressing delayed feedback in conversion rate prediction via influence functions. arXiv preprint arXiv:2502.01669, 2025

  16. [24]

    Fine-tuning pretrained language models: Weight initializations, data orders, and early stopping

    Jesse Dodge, Gabriel Ilharco, Roy Schwartz, Ali Farhadi, Hannaneh Hajishirzi, and Noah Smith. Fine-tuning pretrained language models: Weight initializations, data orders, and early stopping. arXiv preprint arXiv:2002.06305, 2020

  17. [25]

    Who's harry potter? approximate unlearning in llms

    Ronen Eldan and Mark Russinovich. Who's harry potter? approximate unlearning in llms. CoRR, abs/2310.02238, 2023

  18. [26]

    Alphaedit: Null-space constrained knowledge editing for language models

    Junfeng Fang, Houcheng Jiang, Kun Wang, Yunshan Ma, Xiang Wang, Xiangnan He, and Tat - Seng Chua. Alphaedit: Null-space constrained knowledge editing for language models. CoRR, abs/2410.02355, 2024 a

  19. [27]

    Moltc: Towards molecular relational modeling in language models

    Junfeng Fang, Shuai Zhang, Chang Wu, Zhengyi Yang, Zhiyuan Liu, Sihang Li, Kun Wang, Wenjie Du, and Xiang Wang. Moltc: Towards molecular relational modeling in language models. In ACL (Findings) , pp.\ 1943--1958. Association for Computational Linguistics, 2024 b

  20. [28]

    Practical methods of optimization

    Roger Fletcher. Practical methods of optimization. John Wiley & Sons, 2000

  21. [29]

    Maxwell Harper and Joseph A

    F. Maxwell Harper and Joseph A. Konstan. The movielens datasets: History and context. ACM Trans. Interact. Intell. Syst. , 5 0 (4): 0 19:1--19:19, 2016

  22. [30]

    Methods of conjugate gradients for solving linear systems, volume 49

    Magnus Rudolph Hestenes, Eduard Stiefel, et al. Methods of conjugate gradients for solving linear systems, volume 49. NBS Washington, DC, 1952

  23. [31]

    Parameter-efficient transfer learning for NLP

    Neil Houlsby, Andrei Giurgiu, Stanislaw Jastrzebski, Bruna Morrone, Quentin de Laroussilhe, Andrea Gesmundo, Mona Attariyan, and Sylvain Gelly. Parameter-efficient transfer learning for NLP . In ICML , volume 97 of Proceedings of Machine Learning Research, pp.\ 2790--2799. PMLR , 2019

  24. [32]

    Inference without interference: Disaggregate LLM inference for mixed downstream workloads

    Cunchen Hu, Heyang Huang, Liangliang Xu, Xusheng Chen, Jiang Xu, Shuang Chen, Hao Feng, Chenxi Wang, Sa Wang, Yungang Bao, Ninghui Sun, and Yizhou Shan. Inference without interference: Disaggregate LLM inference for mixed downstream workloads. CoRR, abs/2401.11181, 2024 a

  25. [33]

    Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen - Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen

    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. In The Tenth International Conference on Learning Representations, ICLR 2022, Virtual Event, April 25-29, 2022 ....

  26. [34]

    Enhancing sequential recommendation via llm-based semantic embedding learning

    Jun Hu, Wenwen Xia, Xiaolu Zhang, Chilin Fu, Weichang Wu, Zhaoxin Huan, Ang Li, Zuoli Tang, and Jun Zhou. Enhancing sequential recommendation via llm-based semantic embedding learning. In WWW (Companion Volume) , pp.\ 103--111. ACM , 2024 b

  27. [35]

    Exact and efficient unlearning for large language model-based recommendation

    Zhiyu Hu, Yang Zhang, Minghao Xiao, Wenjie Wang, Fuli Feng, and Xiangnan He. Exact and efficient unlearning for large language model-based recommendation. CoRR, abs/2404.10327, 2024 c

  28. [36]

    Knowledge unlearning for mitigating privacy risks in language models

    Joel Jang, Dongkeun Yoon, Sohee Yang, Sungmin Cha, Moontae Lee, Lajanugen Logeswaran, and Minjoon Seo. Knowledge unlearning for mitigating privacy risks in language models. In ACL (1) , pp.\ 14389--14408. Association for Computational Linguistics, 2023

  29. [37]

    LLM maybe longlm: Self-extend LLM context window without tuning

    Hongye Jin, Xiaotian Han, Jingfeng Yang, Zhimeng Jiang, Zirui Liu, Chia - Yuan Chang, Huiyuan Chen, and Xia Hu. LLM maybe longlm: Self-extend LLM context window without tuning. CoRR, abs/2401.01325, 2024

  30. [38]

    Fairsisa: Ensemble post-processing to improve fairness of unlearning in llms

    Swanand Ravindra Kadhe, Anisa Halimi, Ambrish Rawat, and Nathalie Baracaldo. Fairsisa: Ensemble post-processing to improve fairness of unlearning in llms. CoRR, abs/2312.07420, 2023

  31. [39]

    Kassem, Omar Mahmoud, and Sherif Saad

    Aly M. Kassem, Omar Mahmoud, and Sherif Saad. Preserving privacy through dememorization: An unlearning technique for mitigating memorization risks in language models. In EMNLP , pp.\ 4360--4379. Association for Computational Linguistics, 2023

  32. [40]

    Understanding black-box predictions via influence functions

    Pang Wei Koh and Percy Liang. Understanding black-box predictions via influence functions. In ICML , volume 70 of Proceedings of Machine Learning Research, pp.\ 1885--1894. PMLR , 2017

  33. [41]

    Customizing language models with instance-wise lora for sequential recommendation

    Xiaoyu Kong, Jiancan Wu, An Zhang, Leheng Sheng, Hui Lin, Xiang Wang, and Xiangnan He. Customizing language models with instance-wise lora for sequential recommendation. In NeurIPS, 2024

  34. [42]

    Privacy adhering machine un-learning in NLP

    Vinayshekhar Bannihatti Kumar, Rashmi Gangadharaiah, and Dan Roth. Privacy adhering machine un-learning in NLP . In IJCNLP (Findings) , pp.\ 268--277. Association for Computational Linguistics, 2023

  35. [43]

    Towards unbounded machine unlearning

    Meghdad Kurmanji, Peter Triantafillou, Jamie Hayes, and Eleni Triantafillou. Towards unbounded machine unlearning. In NeurIPS, 2023

  36. [44]

    Let machines unlearn - machine unlearning and the right to be forgotten

    Chanhee Kwak, Junyeong Lee, Kyuhong Park, and Heeseok Lee. Let machines unlearn - machine unlearning and the right to be forgotten. In AMCIS . Association for Information Systems, 2017

  37. [45]

    Collavo: Crayon large language and vision model

    Byung - Kwan Lee, Beomchan Park, Chae Won Kim, and Yong Man Ro. Collavo: Crayon large language and vision model. In ACL (Findings) , pp.\ 1121--1138. Association for Computational Linguistics, 2024

  38. [46]

    Block-wise lora: Revisiting fine-grained lora for effective personalization and stylization in text-to-image generation

    Likun Li, Haoqi Zeng, Changpeng Yang, Haozhe Jia, and Di Xu. Block-wise lora: Revisiting fine-grained lora for effective personalization and stylization in text-to-image generation. CoRR, abs/2403.07500, 2024

  39. [47]

    Diffstyler: Diffusion-based localized image style transfer

    Shaoxu Li. Diffstyler: Diffusion-based localized image style transfer. CoRR, abs/2403.18461, 2024

  40. [48]

    Prefix-tuning: Optimizing continuous prompts for generation

    Xiang Lisa Li and Percy Liang. Prefix-tuning: Optimizing continuous prompts for generation. In ACL/IJCNLP (1) , pp.\ 4582--4597. Association for Computational Linguistics, 2021

  41. [49]

    Label supervised llama finetuning

    Zongxi Li, Xianming Li, Yuzhang Liu, Haoran Xie, Jing Li, Fu Lee Wang, Qing Li, and Xiaoqin Zhong. Label supervised llama finetuning. CoRR, abs/2310.01208, 2023

  42. [50]

    Llara: Large language-recommendation assistant

    Jiayi Liao, Sihang Li, Zhengyi Yang, Jiancan Wu, Yancheng Yuan, Xiang Wang, and Xiangnan He. Llara: Large language-recommendation assistant. In SIGIR , pp.\ 1785--1795. ACM , 2024

  43. [51]

    Improved baselines with visual instruction tuning

    Haotian Liu, Chunyuan Li, Yuheng Li, and Yong Jae Lee. Improved baselines with visual instruction tuning. CoRR, abs/2310.03744, 2023 a

  44. [52]

    Pre-train, prompt, and predict: A systematic survey of prompting methods in natural language processing

    Pengfei Liu, Weizhe Yuan, Jinlan Fu, Zhengbao Jiang, Hiroaki Hayashi, and Graham Neubig. Pre-train, prompt, and predict: A systematic survey of prompting methods in natural language processing. ACM Comput. Surv. , 55 0 (9): 0 195:1--195:35, 2023 b

  45. [53]

    Varshney, Mohit Bansal, Sanmi Koyejo, and Yang Liu

    Sijia Liu, Yuanshun Yao, Jinghan Jia, Stephen Casper, Nathalie Baracaldo, Peter Hase, Xiaojun Xu, Yuguang Yao, Hang Li, Kush R. Varshney, Mohit Bansal, Sanmi Koyejo, and Yang Liu. Rethinking machine unlearning for large language models. CoRR, abs/2402.08787, 2024 a

  46. [54]

    P-tuning v2: Prompt tuning can be comparable to fine-tuning universally across scales and tasks

    Xiao Liu, Kaixuan Ji, Yicheng Fu, Zhengxiao Du, Zhilin Yang, and Jie Tang. P-tuning v2: Prompt tuning can be comparable to fine-tuning universally across scales and tasks. CoRR, abs/2110.07602, 2021

  47. [55]

    Understanding llms: A comprehensive overview from training to inference

    Yiheng Liu, Hao He, Tianle Han, Xu Zhang, Mengyuan Liu, Jiaming Tian, Yutong Zhang, Jiaqi Wang, Xiaohui Gao, Tianyang Zhong, Yi Pan, Shaochen Xu, Zihao Wu, Zhengliang Liu, Xin Zhang, Shu Zhang, Xintao Hu, Tuo Zhang, Ning Qiang, Tianming Liu, and Bao Ge. Understanding llms: A c...

  48. [56]

    Towards safer large language models through machine unlearning

    Zheyuan Liu, Guangyao Dou, Zhaoxuan Tan, Yijun Tian, and Meng Jiang. Towards safer large language models through machine unlearning. In ACL (Findings) , pp.\ 1817--1829. Association for Computational Linguistics, 2024 c

  49. [57]

    Eraser: Jailbreaking defense in large language models via unlearning harmful knowledge

    Weikai Lu, Ziqian Zeng, Jianwei Wang, Zhengdong Lu, Zelin Chen, Huiping Zhuang, and Cen Chen. Eraser: Jailbreaking defense in large language models via unlearning harmful knowledge. CoRR, abs/2404.05880, 2024

  50. [58]

    QUARK: controllable text generation with reinforced unlearning

    Ximing Lu, Sean Welleck, Jack Hessel, Liwei Jiang, Lianhui Qin, Peter West, Prithviraj Ammanabrolu, and Yejin Choi. QUARK: controllable text generation with reinforced unlearning. In NeurIPS, 2022

  51. [59]

    Linear and nonlinear programming, volume 2

    David G Luenberger, Yinyu Ye, et al. Linear and nonlinear programming, volume 2. Springer, 1984

  52. [60]

    Llm-rec: Personalized recommendation via prompting large language models

    Hanjia Lyu, Song Jiang, Hanqing Zeng, Yinglong Xia, Qifan Wang, Si Zhang, Ren Chen, Christopher Leung, Jiajie Tang, and Jiebo Luo. Llm-rec: Personalized recommendation via prompting large language models. In NAACL-HLT (Findings) , pp.\ 583--612. Association for Computational L...

  53. [61]

    Reinforced prompt personalization for recommendation with large language models

    Wenyu Mao, Jiancan Wu, Weijian Chen, Chongming Gao, Xiang Wang, and Xiangnan He. Reinforced prompt personalization for recommendation with large language models. CoRR, abs/2407.17115, 2024

  54. [62]

    Getting to know your neighbors (KYN)

    Joanna Misztal - Radecka and Bipin Indurkhya. Getting to know your neighbors (KYN). explaining item similarity in nearest neighbors collaborative filtering recommendations. In UMAP (Adjunct Publication) , pp.\ 59--64. ACM , 2020

  55. [63]

    Feature unlearning for pre-trained gans and vaes

    Saemi Moon, Seunghyuk Cho, and Dongwoo Kim. Feature unlearning for pre-trained gans and vaes. In AAAI , pp.\ 21420--21428. AAAI Press, 2024

  56. [64]

    Hellendoorn, Bogdan Vasilescu, and Brad A

    Daye Nam, Andrew Macvean, Vincent J. Hellendoorn, Bogdan Vasilescu, and Brad A. Myers. Using an LLM to help with code understanding. In ICSE , pp.\ 97:1--97:13. ACM , 2024

  57. [65]

    Nesterov

    Yurii E. Nesterov. Gradient methods for minimizing composite functions. Math. Program., 140 0 (1): 0 125--161, 2013

  58. [66]

    Numerical optimization

    Jorge Nocedal and Stephen J Wright. Numerical optimization. Springer, 1999

  59. [67]

    Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll L. Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, John Schulman, Jacob Hilton, Fraser Kelton, Luke Miller, Maddie Simens, Amanda Askell, Peter Welinder, Paul F. Christiano, Jan Leik...

  60. [68]

    In-context unlearning: Language models as few shot unlearners

    Martin Pawelczyk, Seth Neel, and Himabindu Lakkaraju. In-context unlearning: Language models as few shot unlearners. arXiv preprint arXiv:2310.07579, 2023

  61. [69]

    Pazzani and Daniel Billsus

    Michael J. Pazzani and Daniel Billsus. Content-based recommendation systems. In The Adaptive Web, volume 4321 of Lecture Notes in Computer Science, pp.\ 325--341. Springer, 2007

  62. [70]

    Pearlmutter

    Barak A. Pearlmutter. Fast exact multiplication by the hessian. Neural Comput., 6 0 (1): 0 147--160, 1994

  63. [71]

    The frontier of data erasure: Machine unlearning for large language models

    Youyang Qu, Ming Ding, Nan Sun, Kanchana Thilakarathna, Tianqing Zhu, and Dusit Niyato. The frontier of data erasure: Machine unlearning for large language models. CoRR, abs/2403.15779, 2024

  64. [72]

    Language models are unsupervised multitask learners

    Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. Language models are unsupervised multitask learners. OpenAI blog, 1 0 (8): 0 9, 2019

  65. [73]

    Victor Sanh, Albert Webson, Colin Raffel, Stephen H. Bach, Lintang Sutawika, Zaid Alyafeai, Antoine Chaffin, Arnaud Stiegler, Arun Raja, Manan Dey, M Saiful Bari, Canwen Xu, Urmish Thakker, Shanya Sharma Sharma, Eliza Szczechla, Taewoon Kim, Gunjan Chhablani, Nihal V. Nayak, D...

  66. [74]

    Remember what you want to forget: Algorithms for machine unlearning

    Ayush Sekhari, Jayadev Acharya, Gautam Kamath, and Ananda Theertha Suresh. Remember what you want to forget: Algorithms for machine unlearning. In NeurIPS, pp.\ 18075--18086, 2021

  67. [75]

    Understanding Machine Learning - From Theory to Algorithms

    Shai Shalev - Shwartz and Shai Ben - David. Understanding Machine Learning - From Theory to Algorithms. Cambridge University Press, 2014

  68. [76]

    A survey of large language models on generative graph analytics: Query, learning, and applications

    Wenbo Shang and Xin Huang. A survey of large language models on generative graph analytics: Query, learning, and applications. CoRR, abs/2404.14809, 2024

  69. [77]

    Language models encode collaborative signals in recommendation

    Leheng Sheng, An Zhang, Yi Zhang, Yuxin Chen, Xiang Wang, and Tat-Seng Chua. Language models encode collaborative signals in recommendation. arXiv preprint arXiv:2407.05441, 2024

  70. [78]

    An introduction to the conjugate gradient method without the agonizing pain

    Jonathan Richard Shewchuk et al. An introduction to the conjugate gradient method without the agonizing pain. 1994

  71. [79]

    Graphgpt: Graph instruction tuning for large language models

    Jiabin Tang, Yuhao Yang, Wei Wei, Lei Shi, Lixin Su, Suqi Cheng, Dawei Yin, and Chao Huang. Graphgpt: Graph instruction tuning for large language models. In SIGIR , pp.\ 491--500. ACM , 2024

  72. [80]

    Llama: Open and efficient foundation language models

    Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie - Anne Lachaux, Timoth \' e e Lacroix, Baptiste Rozi \` e re, Naman Goyal, Eric Hambro, Faisal Azhar, Aur \' e lien Rodriguez, Armand Joulin, Edouard Grave, and Guillaume Lample. Llama: Open and efficient fo...

  73. [81]

    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, Dan Bikel, Lukas Blecher, Cristian Canton - Ferrer, Moya Chen, Guillem Cucurull, David Esiobu, Jude Fernandes, Jeremy Fu,...

  74. [82]

    Statistical learning theory

    Vladimir Vapnik. Statistical learning theory. John Wiley & Sons google schola, 2: 0 831--842, 1998

  75. [83]

    Towards efficient and effective unlearning of large language models for recommendation

    Hangyu Wang, Jianghao Lin, Bo Chen, Yang Yang, Ruiming Tang, Weinan Zhang, and Yong Yu. Towards efficient and effective unlearning of large language models for recommendation. CoRR, abs/2403.03536, 2024

  76. [84]

    Zhao, Kelvin Guu, Adams Wei Yu, Brian Lester, Nan Du, Andrew M

    Jason Wei, Maarten Bosma, Vincent Y. Zhao, Kelvin Guu, Adams Wei Yu, Brian Lester, Nan Du, Andrew M. Dai, and Quoc V. Le. Finetuned language models are zero-shot learners. In ICLR . OpenReview.net, 2022

  77. [85]

    Graph convolution machine for context-aware recommender system

    Jiancan Wu, Xiangnan He, Xiang Wang, Qifan Wang, Weijian Chen, Jianxun Lian, and Xing Xie. Graph convolution machine for context-aware recommender system. Frontiers Comput. Sci., 16 0 (6): 0 166614, 2022

  78. [86]

    GIF: A general graph unlearning strategy via influence function

    Jiancan Wu, Yi Yang, Yuchun Qian, Yongduo Sui, Xiang Wang, and Xiangnan He. GIF: A general graph unlearning strategy via influence function. In WWW , pp.\ 651--661. ACM , 2023

  79. [87]

    On the effectiveness of sampled softmax loss for item recommendation

    Jiancan Wu, Xiang Wang, Xingyu Gao, Jiawei Chen, Hongcheng Fu, and Tianyu Qiu. On the effectiveness of sampled softmax loss for item recommendation. ACM Trans. Inf. Syst. , 42 0 (4): 0 98:1--98:26, 2024 a

  80. [88]

    \( \) -dpo: Direct preference optimization with dynamic \( \)

    Junkang Wu, Yuexiang Xie, Zhengyi Yang, Jiancan Wu, Jinyang Gao, Bolin Ding, Xiang Wang, and Xiangnan He. \( \) -dpo: Direct preference optimization with dynamic \( \) . In NeurIPS, 2024 b

  81. [89]

    Evaluating and analyzing relationship hallucinations in large vision-language models

    Mingrui Wu, Jiayi Ji, Oucheng Huang, Jiale Li, Yuhang Wu, Xiaoshuai Sun, and Rongrong Ji. Evaluating and analyzing relationship hallucinations in large vision-language models. In ICML . OpenReview.net, 2024 c

  82. [90]

    Rethinking reinforcement learning for recommendation: A prompt perspective

    Xin Xin, Tiago Pimentel, Alexandros Karatzoglou, Pengjie Ren, Konstantina Christakopoulou, and Zhaochun Ren. Rethinking reinforcement learning for recommendation: A prompt perspective. In SIGIR , pp.\ 1347--1357. ACM , 2022

  83. [91]

    LLM and GNN are complementary: Distilling LLM for multimodal graph learning

    Junjie Xu, Zongyu Wu, Minhua Lin, Xiang Zhang, and Suhang Wang. LLM and GNN are complementary: Distilling LLM for multimodal graph learning. CoRR, abs/2406.01032, 2024 a

  84. [92]

    Hendler, Marzyeh Ghassemi, Anind K

    Xuhai Xu, Bingsheng Yao, Yuanzhe Dong, Saadia Gabriel, Hong Yu, James A. Hendler, Marzyeh Ghassemi, Anind K. Dey, and Dakuo Wang. Mental-llm: Leveraging large language models for mental health prediction via online text data. Proc. ACM Interact. Mob. Wearable Ubiquitous Techno...

  85. [93]

    Rehg, and Aidong Zhang

    Wenqian Ye, Guangtao Zheng, Yunsheng Ma, Xu Cao, Bolin Lai, James M. Rehg, and Aidong Zhang. Mm-spubench: Towards better understanding of spurious biases in multimodal llms. CoRR, abs/2406.17126, 2024

  86. [94]

    Unlearning bias in language models by partitioning gradients

    Charles Yu, Sullam Jeoung, Anish Kasi, Pengfei Yu, and Heng Ji. Unlearning bias in language models by partitioning gradients. In ACL (Findings) , pp.\ 6032--6048. Association for Computational Linguistics, 2023

  87. [95]

    Bitfit: Simple parameter-efficient fine-tuning for transformer-based masked language-models

    Elad Ben Zaken, Yoav Goldberg, and Shauli Ravfogel. Bitfit: Simple parameter-efficient fine-tuning for transformer-based masked language-models. In ACL (2) , pp.\ 1--9. Association for Computational Linguistics, 2022

  88. [96]

    When scaling meets LLM finetuning: The effect of data, model and finetuning method

    Biao Zhang, Zhongtao Liu, Colin Cherry, and Orhan Firat. When scaling meets LLM finetuning: The effect of data, model and finetuning method. In ICLR . OpenReview.net, 2024 a

  89. [97]

    Adaptive budget allocation for parameter-efficient fine-tuning

    Qingru Zhang, Minshuo Chen, Alexander Bukharin, Pengcheng He, Yu Cheng, Weizhu Chen, and Tuo Zhao. Adaptive budget allocation for parameter-efficient fine-tuning. In ICLR . OpenReview.net, 2023

  90. [98]

    Personalized lora for human-centered text understanding

    You Zhang, Jin Wang, Liang - Chih Yu, Dan Xu, and Xuejie Zhang. Personalized lora for human-centered text understanding. In AAAI , pp.\ 19588--19596. AAAI Press, 2024 b

  91. [99]

    Adam can converge without any modification on update rules

    Yushun Zhang, Congliang Chen, Naichen Shi, Ruoyu Sun, and Zhi - Quan Luo. Adam can converge without any modification on update rules. In NeurIPS, 2022

  92. [100]

    Let me do it for you: Towards LLM empowered recommendation via tool learning

    Yuyue Zhao, Jiancan Wu, Xiang Wang, Wei Tang, Dingxian Wang, and Maarten de Rijke. Let me do it for you: Towards LLM empowered recommendation via tool learning. In SIGIR , pp.\ 1796--1806. ACM , 2024

  93. [101]

    McNee, Joseph A

    Cai - Nicolas Ziegler, Sean M. McNee, Joseph A. Konstan, and Georg Lausen. Improving recommendation lists through topic diversification. In WWW , pp.\ 22--32. ACM , 2005

  94. [102]

    write newline

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

  95. [103]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  96. [104]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@firs...

  97. [105]

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.