betaFeedback

WorksComputer Science

TopoSteer: Topological Layer Selection for Activation Steering in LLMs

Abstract

Activation steering controls large language models at inference time by adding a vector to their internal representations. Most work asks how to build that vector, and far less attention has gone to where to add it, even though steering at a poorly chosen layer can be worse than not steering at all. Prior selectors choose a layer without ever steering the model to test it: they score how separable the target behavior is at that layer, and how consistently the individual examples point along the steering direction. We argue that what matters is not the separability of the two activation clouds, nor the alignment of individual displacements with their mean, but the extent of the cloud of per-example displacements that mean-difference steering averages. We propose a topology-based layer selection method, called TopoSteer, which scores a layer by the total H_0 persistence of its displacement cloud, computed as the length of the cloud's Euclidean minimum spanning tree. We show that the score focuses on the displacement cloud's relative geometry, while class separation affects it only through standardization. We also provide a theoretical explanation for our method by proving that, as example-level variation increases, TopoSteer and discriminability move in opposite directions, so the two criteria can rank layers in opposite ways. Across multiple language models and steering tasks, TopoSteer provides a principled alternative to discriminability- and separability-based layer selection and yields more effective steering without requiring held-out validation.

Introduction

Large language models (LLMs) perform strongly across diverse tasks and are now deployed as commercial chat products whose real-world usage is long-form, multilingual, and highly varied across platforms [33], so deployment increasingly demands controllable, reliable, and aligned behavior at inference time. While fine-tuning and reinforcement learning can improve alignment, they are costly, task-specific, and irreversible. Activation steering has therefore emerged as a lightweight alternative that controls LLMs behavior at inference time by intervening directly on internal representations, without modifying model parameters [29]. The dominant recipe is mean-difference (MD) steering, which computes a steering vector from contrastive examples and injects it into selected layers, and which has shown effectiveness for AI alignment, model reasoning, and safety mitigation [24, 19 b]. However, MD’s steering effectiveness depends strongly on where the intervention is applied [25, 27].

This strong layer dependence makes validation-free layer selection an important practical problem: given the contrastive examples used to construct the steering vector, can we predict effective intervention layers without exhaustively evaluating steering at every candidate layer? Figure 1(a) illustrates the challenge using MD as an example. The effectiveness of MD steering varies substantially across injection layers: on Llama-8B-Instruct, steering at early layers leaves performance near the unsteered baseline, injecting at layer 15 improves it, and at other layers steering reduces performance below the baseline. Exhaustively selecting KK of LL layers requires evaluating (LK)\binom{L}{K} possible layer subsets on held-out data, making such a brute-force search impractical. Recent work therefore scores layers using only the extraction data. LayerNavigator Sun et al. [27] ranks layers by Sl=Dl+ClS_l = D_l + C_l, combining a discriminability term DlD_l, defined as a Fisher-type ratio of between-class to total scatter along the steering direction, with ClC_l, which captures the alignment of individual displacements with the steering direction.

Figure 1. Layer choice decides the outcome, and LayerNavigator and TopoSteer choose differently. (a) Single-layer steering performance against depth (conscientiousness, Llama-8B) with the unsteered baseline, and the layer each method ranks highest (K=1K=1). (b) Why they differ: two layers from the family of Theorem 1, with the same class means and the same steering vector, differing only in the scale of example-level variation. Top: activations, each contrastive pair joined by its displacement δl\delta_l (grey arrows), with the mean difference μδ\mu_\delta in blue. With large variation, the two classes overlap, so discriminability DD drops, while the displacements keep a common direction but vary widely in length. Bottom: the same displacements drawn from a common origin form the displacement cloud Δl\Delta_l; connectivity length L\mathcal{L} is the length of its spanning tree (orange), which by Corollary 1 does not see the offset μδ\mu_\delta at all. The large-variation layer is therefore scored higher by TopoSteer and lower by LayerNavigator, whose full score D+CD+C also drops. All values shown are computed from the data drawn; the construction is the one of Appendix § 17.

The two terms capture complementary but incomplete aspects of the geometry relevant to MD steering. MD does not use the two activation clouds separately. At layer ll, let al(x)∈Rda_l(x) \in\mathbb{R}^d denote the residual-stream activation of input xx. We define the per-example displacement as δil:=al(xi+)−al(xi−)\delta_i^l := a_l(x_i^+) - a_l(x_i^-) (in the standardized representation of § 3.1), and the displacement cloud as Δl:={δil}i=1N\Delta_l := \{\delta_i^l\}_{i=1}^N. Because MD steering forms its intervention by averaging these per-example displacements, whether a single vector represents the available signal is a property of the displacement cloud Δl\Delta_l. The discriminability term DlD_l never looks at that cloud: it is built from class means and within-class scatter, and describes how separable the two classes are rather than how the individual examples move. The consistency term does look at it, but only at directions; as an average of cosine similarities, it is unchanged if each displacement is rescaled independently. It therefore measures how tightly the displacements concentrate about the mean direction, and discards their extent: whether the cloud is a tight clump around a single vector, or a broad set of example-specific ones. This extent, overlooked by both terms, is what we propose to score in this paper.

To this end, we propose TOPOSTEER, which scores each layer ll by the total degree-zero persistence of its displacement cloud

sl:=TotPers⁡0(Δl)=∑e∈MST⁡(Δl)w(e)=:L(Δl),(1)s_l := \operatorname{TotPers}_0(\Delta_l) = \sum_{e \in\operatorname{MST}(\Delta_l)} w(e) =: \mathcal{L}(\Delta_l), \tag*{(1)}

evaluated in the standardized representation of § 3.1. The second equality is the classical identification of finite H0H_0 death scales with Euclidean minimum-spanning-tree edge weights [8], which we use as the computational form of the score. The score requires no held-out steering runs, no auxiliary training, and no forward passes beyond those already performed to extract the MD’s steering vectors.

Our analysis shows that TOPOSTEER measures a quantity neither existing term captures. Against discriminability, the relation is an exact reversal: along the axis that scales example-level variation at fixed class means, the two scores move in opposite directions, so the criteria can rank layers oppositely (Theorem 1). Against consistency, it is complementary: connectivity length is invariant to the mean displacement and 1-homogeneous, while consistency is an average of cosines and therefore reads only directions, discarding the extent that L\mathcal{L} measures. On real activations, we further identify an answer-polarity axis induced by the contrastive format rather than the target behavior. This axis splits the displacement cloud into two lobes without contributing between-class information, yet it affects the two criteria through different geometric mechanisms: its effect on discriminability depends on its alignment with the steering direction, whereas its effect on L\mathcal{L} is governed by the spanning-tree bridge and the merging of the two lobes. We analyze these effects in § 5.3.

Our main contributions are as follows. We identify the displacement cloud as the steering-relevant object for MD layer selection and propose TOPOSTEER, a validation-free topological selector based on its total H0H_0 persistence. We theoretically characterize how its connectivity differs from discriminability and directional consistency, and empirically show that it improves layer selection across multiple models and steering tasks without held-out search.

Related Work

Activation Steering and Layer Selection. Activation steering computes a direction associated with a concept or behavior and adds it to residual-stream representations at inference time [34, 29]. A common approach constructs this direction from differences between paired contrastive activations, including contrastive activation addition (Rimsky et al., 2024b), refusal-direction interventions [1], and related inference-time interventions [14]; the model parameters remain unchanged and the intervention can be applied dynamically [19 b]. Our contribution is orthogonal to the choice of direction: we take the extractor as given and ask where to apply it. Existing selection strategies rely on probing accuracy, linear separability, activation differences, or heuristics such as intervening in the middle third of the network. LayerNavigator [27] is the closest point of comparison, combining a Fisher-type discriminability score DD, computed from class means and within-class scatter, with a directional-consistency term CC, the mean cosine between the steering vector and the individual displacements. TopoSteer scores the extent of those displacements, which neither term captures. TopoSteer evaluates the geometry of the per-example displacement vectors that MD steering actually averages, rather than the separability of the two activation classes.

Persistent Homology of Neural Representations. Persistent homology gives a multiscale description of a point cloud [2], and has been used to study neural representations and their geometry [23, 16, 30] (Gabrielli et al. 2020). Fay et al. (2026) analyze LLMs’ activation spaces and show that adversarial inputs induce topological compression, in which diverse small-scale features collapse into fewer dominant large-scale structures. Related work on transformer geometry has also documented systematic layerwise variation in intrinsic dimension, anisotropy, and representational expansion and compression with depth [7]. Recent work has moved toward adaptive steering, selecting intervention layers or steering strengths based on the input. W2S [6] learns a predictor for the input-specific optimal intervention layer, using candidate-layer evaluations to construct the supervision signal. Dynamically Scaled Activation Steering [4] instead learns input- and layer-dependent scaling factors that determine when and how strongly to steer. Nelwan & Wicaksono (2026) extend adaptive selection to multiple layers, training a prompt-based ranker from per-instance layer-selection targets and an adaptive gate for the number of layers. These approaches require either candidate-layer evaluation or additional training to learn an input-dependent selection policy, whereas TopoSteer selects layers directly from the contrastive extraction data, without held-out steering evaluations or selector training. We therefore use persistent homology not as a classifier or training objective, but as a validation-free diagnostic of the geometry of paired displacements for selecting intervention layers.

TopoSteer: Topological Layer Selection

This section defines the method: the object it scores (§ 3.1), the score and its exact invariances (§ 3.2), and the selection procedure, its cost, the diagnostic that guards its main failure mode, and the two regimes in which the selected layers are steered (§ 3.3). Why this criterion differs from class-separability selectors, and what the displacement cloud looks like in a trained model, are deferred to § 5, after the empirical comparison.

The Displacement Cloud

Let ll be a fixed layer of width dd and let {(xi+,xi−)}i=1N\{(x_i^+, x_i^-)\}_{i=1}^{N} be paired contrastive inputs. Write al(xic)∈Rda_l(x_i^c) \in\mathbb{R}^d, c∈{+,−}c \in\{+, -\}, for the residual-stream activation at the final prompt token. MD vectors are computed from the raw activations as follows: vlraw:=1N∑i=1N[al(xi+)−al(xi−)]v_l^{\mathrm{raw}} := \frac{1}{N}\sum_{i=1}^{N}[a_l(x_i^+) - a_l(x_i^-)] and injects αvlraw\alpha v_l^{\mathrm{raw}} at the chosen layers. For selection, we need displacement geometries that are comparable across layers rather than dominated by coordinate-wise scale, we standardize the representation with

a~l(xic)=al(xic)−μlσl,(2)\tilde{a}_l(x_i^c) = \frac{a_l(x_i^c) - \mu_l}{\sigma_l}, \tag*{(2)}

where μl,σl∈Rd\mu_l,\sigma_l \in\mathbb{R}^d are the coordinate-wise mean and standard deviation over all 2N2N activations at layer ll. The displacement cloud is

δli:=a~l(xi+)−a~l(xi−),Δl:={δli}i=1N,(3)\delta_l^i := \tilde{a}_l(x_i^+) - \tilde{a}_l(x_i^-), \qquad\Delta_l := \{\delta_l^i\}_{i=1}^{N}, \tag*{(3)}

with mean δˉl=μ~l+−μ~l−=:μδ,l\bar{\delta}_l = \tilde{\mu}_l^+ - \tilde{\mu}_l^- =: \mu_{\delta,l} is the mean-difference direction in the analyzed space. Writing εl(xic):=a~l(xic)−μ~lc\varepsilon_l(x_i^c) := \tilde{a}_l(x_i^c) - \tilde{\mu}_l^c for the within-class residuals and ξli:=εl(xi+)−εl(xi−)\xi_l^i := \varepsilon_l(x_i^+) - \varepsilon_l(x_i^-) for the residual displacements, each displacement decomposes as

δli=μδ,l+ξli,∑i=1Nξli=0,Ξl:={ξli}i=1N.(4)\delta_l^i = \mu_{\delta,l} + \xi_l^i, \qquad\sum_{i=1}^{N}\xi_l^i = 0, \qquad\Xi_l := \{\xi_l^i\}_{i=1}^{N}. \tag*{(4)}

The contrastive construction gives Δl\Delta_l a second, structural decomposition that the score turns out to be insensitive to; we defer it to § 5.3, where it is what separates the two criteria.

Scoring a Layer

We score a layer by the connectivity of its displacement cloud, measured by the total degree-zero persistence TotPers⁡(X)\operatorname{TotPers}(X) of a finite point cloud XX, the sum of the finite death scales of its Vietoris–Rips H0H_0 barcode (Appendix § 12). Under the Euclidean filtration, these death scales are the edge weights of the Euclidean minimum spanning tree [8], which we use as the computational form of the score,

TotPers⁡(X)=∑e∈MST⁡(X)w(e)=:L(X),(5)\operatorname{TotPers}(X) = \sum_{e \in\operatorname{MST}(X)} w(e) =: \mathcal{L}(X), \tag*{(5)}

restated as Proposition 2 in Appendix § 13. It can be computed from the pairwise distance matrix in O(n2d)O(n^2d) time, and admits a direct interpretation as the connectivity length of the displacement cloud. TopoSteer assigns layer ll the score:

sl:=TotPers⁡(Δl)=L(Δl).(6)s_l := \operatorname{TotPers}(\Delta_l) = \mathcal{L}(\Delta_l). \tag*{(6)}

Two exact properties of L\mathcal{L} determine what the score can and cannot see (Proposition 4). It is translation invariant, and it is 1-homogeneous: L(X+u)=L(X)\mathcal{L}(X + u) = \mathcal{L}(X) for any u∈Rdu \in\mathbb{R}^d and L(cX)=cL(X)\mathcal{L}(cX) = c\mathcal{L}(X) for c>0c > 0. Applying translation invariance to the decomposition equation 4 gives the property we rely on throughout.

Corollary 1 (Common-mean invariance). Under the Euclidean metric, L(Δl)=L(Δl−μδ,l)=L(Ξl)\mathcal{L}(\Delta_l) = \mathcal{L}(\Delta_l - \mu_{\delta,l}) = \mathcal{L}(\Xi_l). The score is therefore invariant to the common mean displacement of the cloud it is given, and depends only on the relative arrangement of the example-level residual displacements. It does not follow that the pipeline is insensitive to between-class separation: the standardization equation 2 is estimated from the same activations, and separation enters through it. Proposition 1 makes the dependence exact.

Layer Selection and Steering Protocols

Layer Selection. Given a layer budget KK, TopoSteer selects the KK highest-scoring layers, SKS_K per task. At inference time, each selected layer receives its corresponding raw MD vector,

al,t′=al,t+αvlraw,l∈SK,t∈T,(7)a'_{l,t} = a_{l,t} + \alpha v_l^{\mathrm{raw}}, \qquad l \in S_K,\quad t \in\mathcal{T}, \tag*{(7)}

where α\alpha is the steering strength and T\mathcal{T} denotes the intervention token positions. The standardized displacement cloud Δl\Delta_l is used only for layer scoring, while the raw MD vector vlrawv_l^{\mathrm{raw}} is used for intervention; TopoSteer therefore changes only the layer-selection rule. Algorithm 1 summarizes the procedure. The same cached activations produce both the steering vectors and the layer scores, so TopoSteer requires no held-out steering evaluations, no auxiliary selector training, and no forward passes beyond those already needed for MD extraction; its additional cost is an offline geometric computation over the extraction data, O(N2d)O(N^2d) per layer (Appendix H). A large connectivity length can arise from edge weights spread across the cloud or from a few long edges bridging separated modes or isolated outliers; in the second case L\mathcal{L} measures separation between modes rather than the arrangement within them. Algorithm 1 therefore also returns the bridge share

bl:=max⁡e∈MST⁡(Δl)w(e)L(Δl),(8)b_l := \frac{\max_{e \in\operatorname{MST}(\Delta_l)} w(e)}{\mathcal{L}(\Delta_l)}, \tag*{(8)}

the fraction of the score carried by its longest edge, which flags whether the score is bulk-driven or carried by a single connection (Appendix F).

References

  1. [1]Andy Arditi, Oscar Balcells Obeso, Aa quib Syed, Daniel Paleka, Nina Rimsky, Wes Gurnee, and Neel Nanda. Refusal in language models is mediated by a single direction. In The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024. URL https://openreview.net/forum?id=pH3XAQME6c.arxiv.org/abs/2406.11717
  2. [2]Herbert Edelsbrunner and John Harer. Persistent homology: A survey. Contemporary Mathematics, 2008. URL https://pub.ista.ac.at/~edels/Papers/2008-02-PersistentHomology.pdf.DOI
  3. [4]Alex Ferrando, Xavier Suau, Jordi González, and Pau Rodriguez. Dynamically scaled activation steering. arXiv preprint arXiv:2512.03661, 2025. URL https://arxiv.org/abs/2512.03661.
  4. [6]Soham Gadgil, Chris Lin, and Su-In Lee. Where to steer: Input-dependent layer selection for steering improves llm alignment. arXiv preprint arXiv:2604.03867, 2026. URL https://arxiv.org/abs/2604.03867.
  5. [7]Yuri Gardinazzi, Karthik Viswanathan, Giada Panerai, Alessio ansuini, Alberto Cazzaniga, and Matteo Biagetti. Persistent topological features in large language models. In Forty-second International Conference on Machine Learning, 2025. URL https://openreview.net/forum?id=qAHnSKhVsm.arxiv.org/abs/2410.11042
  6. [8]J. C. Gower and G. J. S. Ross. Minimum spanning trees and single linkage cluster analysis. Journal of the Royal Statistical Society. Series C (Applied Statistics), 18(1):54–64, 1969. ISSN 00359254, 14679876. URL http://www.jstor.org/stable/2346439.DOI
  7. [9]Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhari, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Sch elten, Alex Vaughan, et al. The llama 3 herd of models. arXiv, 2024. URL https://arxiv.org/abs/2407.21783.
  8. [10]Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika, Dawn Song, and Jacob Steinhardt. Measuring massive multitask language understanding. In International Conference on Learning Representations, 2021. URL https://openreview.net/forum?id=d7KBjmI3GmQ.
  9. [11]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 International conference on machine learning, pp. 2790–2799. PMLR, 2019.arxiv.org/abs/1902.00751
  10. [12]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 International Conference on Learning Representations, 2022. URL https://openreview.net/forum?id=nZeVKeeFYf9.arxiv.org/abs/2106.09685
  11. [13]Lingkai Kong, Haorui Wang, Wenhao Mu, Yuanqi Du, Yuchen Zhuang, Yifei Zhou, Yue Song, Rongzhi Zhang, Kai Wang, and Chao Zhang. Aligning large language models with representation editing: A control perspective. The Thirty-eighth Annual Conference on Neural Information Processing Systems, 37, 2024. https://openreview.net/forum?id=yTTomSJ sSW.arxiv.org/abs/2406.05954
  12. [14]Kenneth Li, Oam Patel, Fernanda Viégas, Hanspeter Pfister, and Martin Wattenberg. Inference-time intervention: Eliciting truthful answers from a language model. Advances in neural information processing systems, 2023. URL https://openreview.net/forum?id=aLLuYpn83y.arxiv.org/abs/2306.03341
  13. [15]Samuel Marks and Max Tegmark. The geometry of truth: Emergent linear structure in large language model representations of true/false datasets. In First Conference on Language Modeling, 2024. URL https://openreview.net/forum?id=ayjHYjjsk.
  14. [16]Gregory Naitzat, Andrey Zhitnikov, and Lek-Heng Lim. Topology of deep neural networks. Journal of Machine Learning Research, 21(184):1–40, 2020. URL http://jmlr.org/papers/v21/20-345.html.arxiv.org/abs/2004.06093
  15. [17]Muhammad Faishal Adly Nelwan and Alfan Farizki Wicaksono. Deployable per-instance multi-layer activation steering for large language models. arXiv preprint arXiv:2608.08829, 2026. URL https://arxiv.org/abs/2608.08829.
  16. [19]Tuc Nguyen and Thai Le. Atlas: Adaptive test-time latent steering with external verifiers for enhancing llms reasoning. arXiv preprint arXiv:2601.03093, 2026a. URL https://arxiv.org/abs/2601.03093v4.
  17. [20]Tuc Nguyen and Thai Le. Beyond linear activation steering: Invertible latent transformations for controlling llm behavior. The Fortieth Annual Conference on Neural Information Processing Systems, 2026b. URL https://arxiv.org/abs/2606.08454.
  18. [21]Tuc Nguyen, Yifan Hu, and Thai Le. Unraveling interwoven roles of large language models in authorship privacy: Obfuscation, mimicking, and verification. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pp. 14914–14930, 2025. URL https://aclanthology.org/2025.emnlp-main.753.pdf.arxiv.org/abs/2505.14195
  19. [22]Ethan Perez, Sam Ringer, Kamile Lukosiute, Karina Nguyen, Edwin Chen, Scott Heiner, Craig Pettit, Catherine Olsson, Sandipan Kundu, Saurav Kadavath, Andy Jones, Anna Chen, Benjamin Mann, Brian Israel, Bryan Seethor, Cameron McKinnon, Christopher Olah, Da Yan, Daniela Amodei, Dario Amodei, Dawn Drain, Dustin Li, Eli Tran-Johnson, Guro Khundadze, Jackson Kernion, James Landis, Jamie Kerr, Jared Mueller, Jeeyoon Hyun, Joshua Landau, Kamal Ndousse, Landon Goldberg, Liane Lovitt, Martin Lucas, Michael Sellitto, Miranda Zhang, Nearav Kingsland, Nelson Elhage, Nicholas Joseph, Noemi Mercado, Nova DasSarma, Oliver Rausch, Robin Larson, Sam McCandlish, Scott Johnston, Shauna Kravec, Sheer El Showk, Tamera Lanham, Timothy Telleen-Lawton, Tom Brown, Tom Henighan, Tristan Hume, Yuntao Bai, Zac Hatfield-Dodds, Jack Clark, Samuel R. Bowman, Amanda Askell, Roger Grosse, Denny Hernandez, Deep Ganguli, Evan Hubinger, Nicholas Schiefer, and Jared Kaplan. Discovering language model behaviors with model-written evaluations. In Anna Rogers, Jordan Boyd-Graber, and Naoaki Okazaki (eds.), Findings of the Association for Computational Linguistics: ACL 2023, pp. 13387–13434, Toronto, Canada, July 2023. Association for Computational Linguistics. doi: 10.18653/v1/2023.findings-acl.847. URL https://aclanthology.org/2023.findings-acl.847/.DOI
  20. [23]Bastian Rieck, Matteo Togninalli, Christian Bock, Michael Moor, Max Horn, Thomas Gumbsch, and Karsten Borgwardt. Neural persistence: A complexity measure for deep neural networks using algebraic topology. In International Conference on Learning Representations, 2019. URL https://openreview.net/forum?id=ByxkjC5FQ.DOI
  21. [24]Nina Rimsky, Nick Gabrieli, Julian Schulz, Meg Tong, Evan Hubinger, and Alexander Turner. Steering llama 2 via contrastive activation addition. In Lun-Wei Ku, Andre Martins, and Vivek Srikumar (eds.), Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp. 15504–15522, Bangkok, Thailand, August 2024a. Association for Computational Linguistics. doi: 10.18653/v1/2024.acl-long.828. URL https://aclanthology.org/2024.acl-long.828/.DOI
  22. [25]Nina Rimsky, Nick Gabrieli, Julian Schulz, Meg Tong, Evan Hubinger, and Alexander Turner. Steering llama 2 via contrastive activation addition. In Lun-Wei Ku, Andre Martins, and Vivek Srikumar (eds.), Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp. 15504–15522, Bangkok, Thailand, August 2024b. Association for Computational Linguistics. doi: 10.18653/v1/2024.acl-long.828. URL https://aclanthology.org/2024.acl-long.828/.DOI
  23. [26]Pau Rodriguez, Arno Blaas, Michal Klein, Luca Zappella, Nicholas Apostoloff, Xavier Suau, et al. Controlling language and diffusion models by transporting activations. In International Conference on Learning Representations, volume 2025, pp. 89812–89855, 2025.arxiv.org/abs/2410.23054
  24. [27]Hao Sun, Huailiang Peng, Qiong Dai, Xu Bai, and Yanan Cao. Layernavigator: Finding promising intervention layers for efficient activation steering in large language models. In The Thirty-ninth Annual Conference on Neural Information Processing Systems, 2025. URL https://openreview.net/forum?id=wj41M45XQR.DOI
  25. [29]Alexander Matt Turner, Lisa Thiergart, Gavin Leech, David Udell, Juan J Vazquez, Ulisse Mini, and Monte MacDiarmid. Steering language models with activation engineering. arXiv, 2023. URL https://arxiv.org/abs/2308.10248.
  26. [30]Adaku Uchendu and Thai Le. Topological data analysis applications in natural language processing: A survey. SIGKDD Explor. Newsl., 28(1):81–101, June 2026. ISSN 1931-0145. doi: 10.1145/3820356.3820362. URL https://doi.org/10.1145/3820356.3820362.
  27. [32]Shitao Xiao, Zheng Liu, Peitian Zhang, Niklas Muennighoff, Defu Lian, and Jian-Yun Nie. C-pack: Packed resources for general chinese embeddings. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR ’24, pp. 641–649, New York, NY, USA, 2024. Association for Computing Machinery. ISBN 9798400704314. doi: 10.1145/3626772.3657878. URL https://doi.org/10.1145/3626772.3657878.
  28. [33]Yueru Yan, Tuc Nguyen, Bo Su, Melissa Lieffers, and Thai Le. Sharechat: A dataset of chatbot conversations in the wild. arXiv preprint arXiv:2512.17843, 2025. URL https://arxiv.org/abs/2512.17843.
  29. [34]Andy Zou, Long Phan, Alisa Liu, et al. Representation engineering: A top-down approach to ai transparency. arXiv preprint arXiv:2310.01405, 2023. URL https://arxiv.org/abs/2310.01405.

Paper details

vlnc:W_t9nsuvs7
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Computer Science

AI tools

View statement
Looking for related papers…
Contents