Networking and Internet Architecture · arXiv:2610.19072 v1 · submitted 13 Oct 2026 · 37 references
Tracing Knowledge Lineage for Federated Unlearning under Mutual Distillation
The paper says
“An edge server and nearby devices can train language models of different sizes together by mutual distillation. Federated unlearning (FU) and unlearning for large language models (LLMs) execute a removal request on one model.”
more of the abstract
Under mutual distillation, however, the effect of the removed data has been copied into many students, and their uploads teach it back to the corrected teacher. We propose Holder-level Erasure via Replication-weighted Delta-closure (HERD), which decides which holders must be corrected. HERD records a knowledge lineage graph from the transfers that the server already performs and weights each edge by its transfer type. It computes how many times the residual at each holder will be copied and selects the cheapest set of holders whose correction bounds the cumulative residual after the request. A transfer guard and a rejoin rule cover the holders left uncorrected or offline. We prove that partial correction suffices only when the copy number of the lineage graph is below one, that averaging alone never reaches one, and that the correction set remains valid under link outages. On TOFU, HERD leaves 18% to 52% less cumulative residual than the strongest baseline at 6% to 12% of its correction cost.
from the abstract
Pith says
Machine review · deepseek-v4-flash · 15 Oct 2026 · what was read was not recorded · not checked by a person
It can be wrong. Challenge this review or reply to it.
The passes disagreed. Reader: holds up. Skeptic: agrees with the reader. Referee: major revision.
HERD solves a real problem (resurrection under mutual distillation) with a clean formulation and order-of-magnitude cost savings, but the load-bearing independence assumption (Assumption 1) is fragile and the paper offers no robustness analysis. This line is wrong
1 major objection · 5 minor
How often objections from reviews like this one are right: not yet measured
Not checked: figures, code, what the 37 references say.
The paper says, at Sec. IV, Assumption 1
The review did not quote this passage.
Pith says: Objection 1 of 1 · major
Proposition 1 states that the guard brings the spectral radius below one when traits spread weakly within a family, as measured by ν. Yet the paper never reports the fitted values of κ_x, κ_s, and ν, nor the resulting spectral radius ρ(ar{W}) before and after the guard, in the TOFU experiments.
more of objection 1
Without this information, it is unclear whether the condition ρ(ar{W})<1 actually holds in the evaluated settings. If it does not, then the theoretical guarantee for partial correction (Theorem 1) would not apply, and the empirical success would lack a theoretical basis. Please report these values and discuss whether the guard actually achieves ρ(ar{W})<1 for the TOFU federation.
5 minor objections
The paper says, at General
Pith says: minor objection 1 of 5
The paper uses 'copy number' to refer to the spectral radius ρ(ar{W}) and 'copy weight' for φ_j. While the distinction is clear in context, a definitional sentence at the end of Section III-B would help avoid confusion.
The paper says, at Fig. 3
Pith says: minor objection 2 of 5
The caption lists 'score' values that are not explicitly defined in the figure or caption. Please add a brief explanation of what the numbers represent (e.g., copy weight or hop weight) and how they are computed for this illustration.
show 3 more
The paper says, at Table I
Pith says: minor objection 3 of 5
The testbed results use only two clients (one same-family, one cross-family), while the simulation uses 50 clients. The paper should note that the hardware numbers are illustrative and may not reflect the scaling behavior shown in Fig. 7a.
The paper says, at Sec. V-A
Pith says: minor objection 4 of 5
The simulation sets δ as a percentage of uncorrected exposure (Exp). The paper should clarify the relationship between the budget δ in Problem (P) and the exposure reported in Table I (i.e., the absolute residual bound vs. the percentage metric).
The paper says, at Eq. (1)
Pith says: minor objection 5 of 5
The Shannon capacity formula in Eq. (1) assumes no interference; this is a standard simplification but should be stated as an assumption.
What would settle it
Simulate N=50 clients over 20 rounds with mutual distillation. Let each client’s link state follow an AR(1) process on the logit scale (autocorrelation ρ=0.3). Let the fraction of distillation prompts that touch u follow a two-state Markov chain with transition probability 0.8 per round (so excitation correlates over time). Fit ¯W from the marginal means, compute φ and select K via HERD. For 100 random seeds, record the actual cumulative residual after correction and compare its mean and 95th percentile to φ^⊤ r+. If the mean exceeds φ^⊤ r+ by more than 10%, or if the 95th percentile exceeds 1.5×φ^⊤ r+, then the independence assumption is materially consequential for the method’s guarantee.
from the skeptic's reading
How the review went
4 readings by deepseek-v4-flash · 15 Oct 2026
- Reader holds up
- Skeptic agrees
- Editor wrote the verdict above
- Referee major revision: 1 major objection, 5 minor
Not run on this paper: the simulated authors' defence.
Pith did not check this review's objections against the paper.
The review in full
The review in brief, machine-written
Thanks for sending this. The paper addresses an important gap: federated unlearning when the server and devices train by mutual distillation. Existing methods either ignore the return path from students back to the teacher (resurrection) or correct all descendants at high cost. HERD formalizes the problem as selecting a minimal-cost set of holders whose correction bounds the cumulative residual, using a typed lineage graph with copy weights that capture whether averaging or distillation spreads the residual.
The core insight—that averaging alone never pushes the spectral radius of the expected inheritance matrix past one, while distillation can—is sound and practically useful. The experiments on TOFU are convincing: HERD leaves 18–52% less cumulative residual than UNDO at 6–12% of its correction cost, across client, class, and sample unlearning scenarios. The theoretical statements (Theorem 1, Propositions 1–3) hold under the stated assumptions.
Now the soft spot. The stress-test concern is real: Assumption 1 (independence of W[t] across rounds) is load-bearing for Theorem 1. The Neumann series argument that E[Σ r[t]] = φᵀ r₊ depends on factoring powers of the mean matrix, which fails when W[t] are correlated. The authors acknowledge that the inheritance matrix is not fixed (sentence [b76]) but do not relax independence. In practice, client participation, link quality, and distillation prompts are correlated over time, so the guarantee might degrade. The paper offers no concentration bound or sensitivity analysis. This is a genuine limitation, not a minor one, because the method's correctness depends on ρ(W̅) < 1 being the right threshold under correlation.
Minor concern: the fitted parameters κₓ, κₛ, ν via canaries introduce calibration dependence, but the authors are upfront about it and it's a practical necessity.
Overall, the paper makes a solid contribution to a problem that matters. The fragility of Assumption 1 should be addressed in revision—either by proving a bound under mixing or by providing empirical robustness checks. I would accept a peer-review invitation and send it to referees, with the independence assumption as the primary discussion point. The work is worth engaging.
The strongest objection
The entire theoretical guarantee that partial correction suffices when ρ(¯W)<1 rests on Assumption 1: the inheritance matrices W[t] are independent across rounds and independent of r[t]. Theorem 1 then uses the Neumann series of ¯W to equate the expected cumulative residual to φ^⊤ r+. If W[t] are not independent, the expectation of the infinite product Π_{s≥t_u} W[s] does not factor into powers of ¯W, and Eq. (8) is no longer guaranteed to bound even the expected cumulative residual. The paper acknowledges that the inheritance matrix is not fixed (sentence [b76]) but does not weaken the independence condition. Real federated systems exhibit correlated client participation (e.g., bursty link outages) and correlated distillation prompts, making the independence assumption fragile. A practical failure would be that selected δ-closure leaves more residual than predicted, or that the cumulative residual exceeds δ even when ρ(¯W)<1. This is a genuine soft spot because the method’s correctness depends on it, and the paper offers no robustness analysis or concentration bound.
The reader’s weakest_assumption correctly identifies Assumption 1 as the most fragile point. The stress-test confirms that this is the single most load-bearing concern: without independence, the copy weights derived from the Neumann series are not guaranteed to bound the expected cumulative residual. However, the paper is transparent about this assumption, states it clearly, and builds the theory on it. The method is novel, the experimental results on TOFU show practical improvement, and the authors acknowledge that the inheritance matrix is not fixed. Therefore, the concern does not warrant changing the ACCEPT verdict; rather it motivates a concrete sensitivity check that a follow-up could address. The verdict stays unchanged because the paper delivers its claimed contributions within the stated assumptions.
The referee's report
If the results hold, HERD addresses a genuine gap in federated unlearning: the resurrection of removed knowledge through mutual distillation loops that cannot be solved by correcting only the teacher or all descendants. The idea of using copy weights derived from the server's own records, without sending the removed data to clients, is novel and practical. The theoretical characterization of when partial correction suffices (spectral radius condition) and the proof that averaging alone never reaches a spectral radius of one are clean insights.
Pith's reading of the paper (a machine's summary, not the authors')
Under mutual distillation in federated learning, deleting data from the teacher model is not enough because the knowledge has been copied into student models and will be taught back. The paper proposes HERD, which records a typed knowledge lineage graph from transfers the server already performs and computes a 'copy weight' for each holder—the total future exposure that one unit of residual creates. HERD then selects the cheapest set of holders whose correction keeps the cumulative residual below a threshold. The central theoretical result is that partial correction suffices only when the copy number (spectral radius of the expected inheritance matrix) is below one, and that averaging alone never reaches one. Experiments on TOFU show HERD leaves 18% to 52% less cumulative residual than the strongest baseline at 6% to 12% of its correction cost.
What carries the argument. The copy number is the spectral radius ρ(W̄) of the expected inheritance matrix W̄, where each entry w̄ᵢⱼ is the expected fraction of holder j's residual that holder i inherits in one round. The copy weight vector φ = (I − W̄ᵀ)⁻¹ 1 gives the total exposure created by one unit of residual at each holder. The typed lineage graph records edges weighted by transfer type: averaging rows sum to one (spreads residual), distillation rows sum to more than one (copies residual). The selection problem is a covering knapsack: minimize correction cost subject to Σⱼ φⱼ rⱼ[tᵤ] ≤ δ for the uncorrected holders.
Core claim. The paper identifies and formalizes the 'resurrection' problem in federated unlearning under mutual distillation: after correcting the teacher, student uploads still carry the removed data's effect and teach it back. To solve this, HERD records every transfer's type (averaging or distillation) into a lineage graph without extra communication, then computes copy weights that account for the entire forward pathway including return edges. The core condition is that partial correction is viable only when the copy number ρ(W̄) < 1; if ρ(W̄) ≥ 1, any nonzero residual leads to unbounded cumulative exposure. The paper further proves that averaging alone never reaches ρ = 1, so distillation is the sole cause of resurrection.
The premise it rests on. Assumption 1 requires that the inheritance matrices W[t] are independent across rounds and independent of r[t], and that their mean W̄ is nonnegative and irreducible; in real federated systems where client participation, link quality, and distillation prompts correlate over time, this independence may fail and irreducibility may hold only in expectation over long periods.
If the paper is right (Pith's inferences, not the authors')
Federated unlearning systems that use mutual distillation must account for the full distillation loop, not just the teacher model, otherwise the removed data resurfaces within a few rounds.
The copy number condition ρ < 1 provides a simple necessary and sufficient test for whether partial correction is feasible; systems can monitor this number online to decide their unlearning strategy.
Averaging alone cannot cause resurrection because every row of an averaging matrix sums to one, keeping the spectral radius below one after scaling by the retention factor λⱼ < 1.
The transfer guard (dropping outputs that carry u) can reduce ρ below one even when distillation edges would otherwise push it above one, but the floor ν of same-family distillation may still allow weak copying.
The rejoin rule for offline holders uses the round of their last transfer to estimate residual, and corrects them before their first upload if their copy weight exceeds the remaining budget, keeping the closure valid.
The paper's independence assumption on inheritance matrices (Assumption 1) may be violated in real systems where client participation and link quality correlate over time; a natural extension would be to derive bounds under Markovian or adversarial perturbations.
Because the copy weight φ = (I − W̄ᵀ)⁻¹ 1 depends on the entire graph, a single student with high out-degree (e.g., a strong same-family absorber) can force the closure to include many holders; scaling analysis for large networks would be valuable.
The technique of using canaries to measure absorption rates κₓ, κₛ, and ν is clever but assumes canaries are representative of u; a robustness check against canary mismatch would strengthen the practical deployment.
The method's applicability extends beyond LLMs to any federated system with heterogeneous models exchanging soft targets, such as vision transformers or recommendation models, as long as the transfers pass through a central server.
Check this review
Pith stored no record of this review's passes and took no fingerprint of it. Reviews made by code from 13 Nov 2026 on carry both.
How often reviews like this one are right
Reviews like this one: written by deepseek-v4-flash, where what was read was not recorded.
How often objections from reviews like this one are right: not yet measured
Reviews by people
This review was written by a machine. No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.
The paper
Abstract
An edge server and nearby devices can train language models of different sizes together by mutual distillation. Federated unlearning (FU) and unlearning for large language models (LLMs) execute a removal request on one model. Under mutual distillation, however, the effect of the removed data has been copied into many students, and their uploads teach it back to the corrected teacher. We propose Holder-level Erasure via Replication-weighted Delta-closure (HERD), which decides which holders must be corrected. HERD records a knowledge lineage graph from the transfers that the server already performs and weights each edge by its transfer type. It computes how many times the residual at each holder will be copied and selects the cheapest set of holders whose correction bounds the cumulative residual after the request. A transfer guard and a rejoin rule cover the holders left uncorrected or offline. We prove that partial correction suffices only when the copy number of the lineage graph is below one, that averaging alone never reaches one, and that the correction set remains valid under link outages. On TOFU, HERD leaves 18% to 52% less cumulative residual than the strongest baseline at 6% to 12% of its correction cost.
The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. No sentence in this record was written by a model.
abstract compiled · 2026-10-15
The authors' main claim, as the classifier picked it out 0.91
“On TOFU, HERD leaves 18% to 52% less cumulative residual than the strongest baseline at 6% to 12% of its correction cost.”
How the main claim is stated: asserted 1.00
What the abstract itself states, as probabilities: claims first 0.05 · states limitation 0.66 · human participants 0.05 · controlled comparison 0.91 · randomized assignment 0.02 · states quantitative result 0.99 · compares with prior results 0.98 · mentions code or data availability 0.03
The abstract, sentence by sentence
This table labels the abstract only (arXiv listing abstract v1). The number is the probability the classifier gave the label.
| Sentence | Role | The authors' own work | States the main conclusion | Strength |
|---|---|---|---|---|
| An edge server and nearby devices can train language models of different sizes together by mutual distillation. | background 1.00 | 0.12 | 0.02 | asserted 0.59 |
| Federated unlearning (FU) and unlearning for large language models (LLMs) execute a removal request on one model. | background 1.00 | 0.07 | 0.03 | asserted 1.00 |
| Under mutual distillation, however, the effect of the removed data has been copied into many students, and their uploads teach it back to the corrected teacher. | background 0.59 | 0.20 | 0.06 | asserted 1.00 |
| We propose Holder-level Erasure via Replication-weighted Delta-closure (HERD), which decides which holders must be corrected. | contribution 0.93 | 0.87 | 0.18 | asserted 1.00 |
| HERD records a knowledge lineage graph from the transfers that the server already performs and weights each edge by its transfer type. | method 0.96 | 0.79 | 0.05 | asserted 1.00 |
| It computes how many times the residual at each holder will be copied and selects the cheapest set of holders whose correction bounds the cumulative residual after the request. | method 0.99 | 0.90 | 0.13 | asserted 1.00 |
| A transfer guard and a rejoin rule cover the holders left uncorrected or offline. | method 0.84 | 0.82 | 0.04 | asserted 1.00 |
| We prove that partial correction suffices only when the copy number of the lineage graph is below one, that averaging alone never reaches one, and that the correction set remains valid under link outages. | contribution 0.63 | 0.93 | 0.30 | asserted 1.00 |
| On TOFU, HERD leaves 18% to 52% less cumulative residual than the strongest baseline at 6% to 12% of its correction cost. | result 1.00 | 0.95 | 0.44 | asserted 1.00 |
What kind of paper the classifier reports this is: method or tool 1.00 · evidence from observation or experiment 0.53
What was examined, and what was not
Sources read
- arXiv listing abstract v1
Not examined
- full text
- figures and tables
- references and citing sentences
- relations to other papers
Levels run
- P abstract
Not run
- S sections
- T body sentences
- C citations
- R cross-paper relations
How the labels were made
Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-10-15: 1 request, 6,881 input tokens.
Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. This record as JSON
Figures (7)
Showing 5 of the paper's 7 figures. Figures are the authors', reproduced from the paper.
References (37)
- [1] B. Wu, Z. Ding, and J. Huang, “A Review of Continual Learning in Edge AI,”IEEE Transactions on Network Science and Engineering, vol. 13, pp. 6571–6588, 2026 on Pith
- [2] Z. Lin, G. Qu, Q. Chen, X. Chen, Z. Chen, and K. Huang, “Pushing Large Language Models to the 6G Edge: Vision, Challenges, and Opportunities,”IEEE Communications Magazine, vol. 63, no. 9, pp. 52–59, 2025 on Pith
- [3] B. Wu, J. Huang, and Y . Zhao, “REBASE: Device-cloud experi- ence coherence for GUI agents across app updates,”arXiv preprint arXiv:2609.32130, 2026 on PithPith's reviewarXiv
- [4] B. Wu, Z. Ding, and J. Huang, “Not all experience belongs in the weights: Component routing for self-improving GUI agents,”arXiv preprint arXiv:2610.01787, 2026 on PithPith's reviewarXiv
- [5] B. Wu and J. Huang, “CrystalMem: Elastic memory for self- evolving LLM agents via knowledge crystallization,”arXiv preprint arXiv:2608.00303, 2026 on PithPith's reviewarXiv
- [6] J. Huang, C.-C. Xing, and M. Guizani, “Power allocation for D2D communications with SWIPT,”IEEE Transactions on Wireless Com- munications, vol. 19, no. 4, pp. 2308–2320, 2020 on Pith
- [7] J. Huang, Y . Zhou, Z. Ning, and H. Gharavi, “Wireless power transfer and energy harvesting: Current status and future prospects,”IEEE Wireless Communications, vol. 26, no. 4, pp. 163–169, 2019 on Pith
- [8] B. Wu, J. Huang, and S. Yu, ““X of Information” Continuum: A Survey on AI-Driven Multi-Dimensional Metrics for Next-Generation Net- worked Systems,”IEEE Communications Surveys & Tutorials, vol. 28, pp. 5307–5344, 2026 on Pith
- [9] B. Wu, Z. Ding, and J. Huang, “RELIEF: Turning missing modalities into training acceleration for federated learning on heterogeneous IoT edge,”IEEE Internet of Things Journal, 2026, early access on Pith
- [10] B. Wu, J. Huang, Q. Duan, L. Dong, and Z. Cai, “Enhancing Vehic- ular Platooning With Wireless Federated Learning: A Resource-Aware Control Framework,”IEEE Transactions on Networking, vol. 34, pp. 1479–1494, 2026 on Pith
- [11] B. Wu, J. Huang, and Q. Duan, “Real-time intelligent healthcare enabled by federated digital twins with AoI optimization,”IEEE Network, vol. 40, no. 2, pp. 184–191, 2026 on Pith
- [12] Z. Ding, J. Huang, Q. Duan, C. Zhang, Y . Zhao, and S. Gu, “A Dual- Level Game-Theoretic Approach for Collaborative Learning in UA V- Assisted Heterogeneous Vehicle Networks,” in2025 IEEE International Performance, Computing, and Communications Conference (IPCCC), 2025, pp. 1–8 on Pith
- [13] Z. Ding, J. Huang, Y . Zhao, and Z. Cai, “Combating knowledge diversity and catastrophic forgetting in UA V-assisted collaborative ve- hicular learning: A game-theoretic approach,”ACM Transactions on Autonomous and Adaptive Systems, Jun. 2026, just Accepted on Pith
- [14] X. Che, D. Xu, and X. Du, “Comparative analysis of formula and structure prediction from tandem mass spectra,” inData Science: Foundations and Applications (PAKDD 2026), ser. Lecture Notes in Computer Science, vol. 16618. Singapore: Springer, 2027, pp. 215–227 on Pith
- [15] S. Itahara, T. Nishio, Y . Koda, M. Morikura, and K. Ya- mamoto, “Distillation-Based Semi-Supervised Federated Learning for Communication-Efficient Collaborative Training With Non-IID Private Data,”IEEE Transactions on Mobile Computing, vol. 22, no. 1, pp. 191–205, 2023 on Pith
- [16] C. He, M. Annavaram, and S. Avestimehr, “Group Knowledge Transfer: Federated Learning of Large CNNs at the Edge,” inAdvances in Neural Information Processing Systems, vol. 33, 2020, pp. 14 068–14 080 on Pith
- [17] T. Fan, G. Ma, Y . Kang, H. Gu, Y . Song, L. Fan, K. Chen, and Q. Yang, “FedMKT: Federated Mutual Knowledge Transfer for Large and Small Language Models,” inProceedings of the 31st International Conference on Computational Linguistics, 2025, pp. 243–255 on Pith
- [18] B. Wu, J. Huang, and Y . Zhao, “From Alpha to Omega: Lifecycle- aware forgetting defense in federated continual learning for planetary exploration,” in2026 IEEE 46th International Conference on Distributed Computing Systems (ICDCS), 2026, pp. 1–11 on Pith
- [19] B. Wu and J. Huang, “Lifecycle-Aware Federated Continual Learning in Mobile Autonomous Systems,” arXiv preprint arXiv:2604.20745, 2026 on PithPith's reviewarXiv
- [20] B. Wu, Z. Ding, and J. Huang, “PRISM: Exposing and Resolving Spurious Isolation in Federated Multimodal Continual Learning,” arXiv preprint arXiv:2605.01061, 2026 on PithPith's reviewarXiv
show 17 more
- [21] Z. Ding, B. Wu, and J. Huang, “Scale: Sensitivity-Aware Federated Unlearning with Information Freshness Optimization for Mobile Edge Computing,” inProceedings of the IEEE 46th International Conference on Distributed Computing Systems (ICDCS), 2026, pp. 1476–1486 on Pith
- [22] H. Wu, Y . Tao, A. Zhou, S. Chen, F. Dressler, and D. Yu, “Certified Unlearning in Decentralized Federated Learning,” inProceedings of the IEEE International Conference on Computer Communications, 2026, pp. 1–10 on Pith
- [23] Z. Ding and J. Huang, “Toward Trustworthy Federated Unlearning for Mobile Autonomous Systems,”IEEE Network, pp. 1–9, 2026 on Pith
- [24] U. Pudasaini, Z. Ding, and J. Huang, “Securing smart agriculture with communication-efficient federated unlearning,” in2026 IEEE 27th International Conference on High Performance Switching and Routing (HPSR), 2026, pp. 1–7 on Pith
- [25] Z. Ding, B. Wu, and J. Huang, “EASE: Federated Multimodal Un- learning via Entanglement-Aware Anchor Closure,” arXiv preprint arXiv:2605.00733, 2026 on PithPith's reviewarXiv
- [26] Z. Ding, J. Huang, and L. Dong, “When unlearning fails: Reliable data deletion under post-training in agent networks,”arXiv preprint arXiv:2607.28829, 2026 on PithPith's reviewarXiv
- [27] C. Fan, J. Liu, L. Lin, J. Jia, R. Zhang, S. Mei, and S. Liu, “Sim- plicity Prevails: Rethinking Negative Preference Optimization for LLM Unlearning,” inAdvances in Neural Information Processing Systems, vol. 38, 2025, pp. 1776–1803 on Pith
- [28] X. Che, D. Xu, and X. Wu, “StraightDP: Geometry-aware dif- ferential privacy for rectified-flow transformers,”arXiv preprint arXiv:2607.29100, 2026 on PithPith's reviewarXiv
- [29] Z. Ding, J. Huang, and J. Qi, “Learning to defend: A multi-agent reinforcement learning framework for Stackelberg security game in mo- bile edge computing,” in2026 International Conference on Computing, Networking and Communications (ICNC), 2026, pp. 769–774 on Pith
- [30] X. Liu, M. Li, G. Yu, L. Li, H. Peng, and R. P. Liu, “Parallel Unlearn- ing in Inherited Model Networks,”IEEE Transactions on Information Forensics and Security, vol. 20, pp. 12 199–12 214, 2025 on Pith
- [31] Y . Quan, Z. Li, and G. Montana, “Efficient Verified Unlearning For Distillation,” inAdvances in Neural Information Processing Systems, vol. 38, 2025, pp. 79 212–79 239 on Pith
- [32] B. W. Lee, A. Foote, A. Infanger, L. Shor, H. Kamath, J. Goldman- Wetzler, B. Woodworth, A. Cloud, and A. M. Turner, “Distillation Robustifies Unlearning,” inAdvances in Neural Information Processing Systems, vol. 38, 2025, pp. 98 457–98 491 on Pith
- [33] A. Cloud, M. Le, J. Chua, J. Betley, A. Sztyber-Betley, S. Minder- mann, J. Hilton, S. Marks, and O. Evans, “Language models transmit behavioural traits through hidden signals in data,”Nature, vol. 652, no. 8110, pp. 615–621, 2026 on Pith
- [34] 3GPP, “Further advancements for E-UTRA physical layer aspects,” 3rd Generation Partnership Project (3GPP), Technical Report TR 36.814, V9.0.0, Mar. 2010 on Pith
- [35] A. Yanget al., “Qwen2.5 Technical Report,” arXiv preprint arXiv:2412.15115, 2024 on PithPith's reviewarXiv
- [36] E. J. Hu, Y . Shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, and W. Chen, “LoRA: Low-Rank Adaptation of Large Language Models,” inInternational Conference on Learning Representations (ICLR), 2022 on Pith
- [37] P. Maini, Z. Feng, A. Schwarzschild, Z. C. Lipton, and J. Z. Kolter, “TOFU: A Task of Fictitious Unlearning for LLMs,” inFirst Conference on Language Modeling (COLM), 2024 on Pith
Challenges and discussion
Disagree with the review, or know something it missed? Leave a comment, or challenge this review formally: a challenge becomes a permanent entry in this paper's record, and its decision is published here.
Come back by email
A short list of the week's newly reviewed papers. Nothing is sent until you confirm from your inbox, and every email has a link that stops it. No account.
No comments yet.
Sign in to comment