Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

TAROT: Targeted Data Selection via Optimal Transport

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read TAROT claims targeted data selection should minimize optimal-transport distance in whitened gradient space, outperforming influence heuristics and full-data training with far less data.

desk verdict Solid empirical paper with a genuine new distance metric, but the OT selection claim is not isolated and the OTM protocol has an internal inconsistency. read the letter →

arxiv 2412.00420 v2 pith:EDOGVJXQ submitted 2024-11-30 cs.LG cs.CVstat.ML

Lan Feng , Fan Nie , Yuejiang Liu , Alexandre Alahi This is my paper · ORCID
classification cs.LGcs.CVstat.ML
keywords targeteddataselectionoptimaltransportwhitenedfeaturedistanceinfluencegradientfeaturesinstructiontuningmotionpredictionsemanticsegmentation
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

TAROT argues that targeted data selection should be posed as distribution matching: pick the subset of a candidate pool whose optimal-transport distance to the target data, measured in a whitened gradient-feature space, is smallest. It identifies two failure modes in influence-based heuristics — dominant feature components swamping high-dimensional distance estimates, and the assumption that per-sample influences add linearly — and replaces them with a whitened-feature distance and a greedy OT-minimizing selection. Across semantic segmentation, motion prediction, and instruction tuning, subsets chosen this way match or beat full-data training while using a fraction of the data; in the LLM setting OTM selects as little as 0.13% of the candidate pool and improves BBH from 63.8 to 68.9. The paper also introduces an early-stopping rule based on when OT distance starts to increase, which estimates the optimal selection ratio without a validation sweep.

What carries the argument

Whitened Feature Distance (WFD): $\phi(z)=\sum_i \nabla L(z;\theta_i)$ summed over checkpoints, projected, decorrelated by Cholesky (or ZCA) whitening, then L2-normalized, giving pair distance $d^w_{\mathcal{Z}}(z,z')=\|\hat\phi(z)-\hat\phi(z')\|_2$. This feeds the OT cost in $d_{OT}(D_s,D_t)=\min_{\pi}\int d^w_{\mathcal{Z}}\,d\pi$, solved by Sinkhorn; the greedy selection uses nearest-neighbor candidates and, for OTM, stops when $d_{OT}$ to held-out folds rises. The machinery does two jobs: it repairs the metric so dominant gradient directions do not dominate selection, and it converts selection into a distribution-alignment problem with a built-in stopping criterion.

What would settle it

Construct a target domain that is a simple unimodal shift of one candidate cluster, with a second candidate cluster of near-duplicate, low-gradient samples placed at large whitened distance: additive influence methods should select the shifted cluster easily, and if TAROT's OT minimization over nearest neighbors still selects fewer useful samples or stops too early (higher OT distance to held-out data than a random subset of the same size), the monotonicity assumption is falsified.

Watch

Extended reading notes

Core claim

The central claim is that the whitened gradient-feature distance (WFD) is a more reliable measure of a sample's usefulness for a target distribution than raw projected-gradient influence, and that minimizing the optimal-transport distance between the selected subset and the target, rather than greedily accumulating high-influence points, yields subsets that generalize better on complex, multimodal target data. TAROT embeds each example by summing loss gradients over checkpoints of a model trained on the candidate pool, projects, whitens, and normalizes to the unit sphere, then uses the resulting Euclidean distances as the OT cost. Its greedy selection iterates nearest neighbors of target points and, in the OTM variant, stops when the OT distance to held-out target folds increases, producing both a subset and an estimated selection ratio. The paper reports consistent gains over LESS, DsDm, TSDS, and full-data training across three task families, including cross-architecture transfer.

Load-bearing premise

The pretrained model's whitened gradient features are assumed to capture, monotonically, how much a training example will help the target task, so that smaller OT distance in this space means better downstream performance.

Editorial extensions

If this is right

  • Training on OT-selected subsets can match or surpass full-dataset training while using a small fraction of data; for example, BBH rises to 68.9 with 0.13% of the candidate pool versus 63.8 with the full dataset.
  • OTM's early-stopping rule estimates an optimal selection ratio automatically, avoiding the budget sweeps that influence-based methods require.
  • WFD outperforms TRAK on the LDS influence-estimation score across CIFAR-10 and nuScenes for every ensemble size tested, indicating a stronger correlation with model behavior.
  • Selection transfers across architectures and model scales (AutoBots to Wayformer, LLaMA to Qwen) and does not require the subtask labels that LESS relies on.

Reading between the lines

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

  • If whitening is the key repair, then applying the same normalization to other embedding spaces, such as feature activations, may extend the method to settings where gradients are unavailable or prohibitively expensive.
  • The early-stopping rule suggests a data-efficiency law: the optimal subset size is where marginal transport cost turns upward, a quantity that could be used to compare candidate data sources before any training.
  • Because OTM selects nearest neighbors in whitened gradient space, it may concentrate on dominant target modes and under-select rare edge cases; this testable risk could be checked by measuring performance on out-of-distribution target splits.
  • Combining WFD with explicit diversity regularization should recover the generalization that the authors themselves flag as a limitation, offering a concrete extension of the current objective.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes TAROT, a targeted data selection framework that uses optimal transport with whitened gradient features. The method first computes per-sample gradient features from a pretrained model, whitens and normalizes them (WFD), then selects subsets either by a fixed-size greedy nearest-neighbor rule or by an OT-distance-minimization rule with k-fold early stopping, and optionally reweights selected samples via OT potentials. Experiments cover semantic segmentation (GTA5 to Cityscapes), motion prediction (multi-source to nuScenes), and instruction tuning (FLAN/CoT/Dolly/OpenAssistant to MMLU/BBH), reporting gains over LESS, DsDm, TSDS, random selection, and full-data training, as well as cross-architecture transfer.

Significance. If the central claim holds, TAROT is a practically valuable framework: it consistently improves targeted data selection across vision and language tasks, can select very small subsets (e.g., 0.13% of the candidate pool for BBH transfer), and exhibits transfer across architectures (Autobots-to-Wayformer and LLaMA-to-Qwen). The paper's strengths include extensive multi-task evaluation, a released codebase, ablations isolating the whitening and normalization components (Table 2), cross-architecture transfer experiments, and computational complexity analysis. However, the load-bearing attribution of the gains to OT-based selection over influence-based heuristics is not fully isolated, and the OTM protocol is internally inconsistent between the main text and the algorithm listing. These issues are fixable with additional experiments and clarification, but they are central to the paper's contribution.

major comments (4)
  1. [§4.2–§4.4 vs. §3.2–§3.3] The central claim that OT-based selection outperforms influence-based additive greedy heuristics is confounded: TAROT changes both the feature/distance (whitened gradient features, Eqs. 6–10) and the selection rule (greedy kNN plus OT potential, Algorithms 1–2), while the LESS and DsDm baselines use unwhitened projected-gradient influence with additive top-k selection. No experiment runs an additive greedy heuristic on the same WFD features or runs TAROT's selection rule on unwhitened features, so the reported gains could be entirely due to WFD rather than to the OT formulation. Please add ablations that isolate the selection rule, e.g., top-k additive selection with WFD cost, or OT-potential selection with the original projected-gradient distance, at least on one segmentation and one instruction-tuning setting.
  2. [§3.3, Eq. (14) and Algorithm 2] There is a direct inconsistency between the stopping-criterion description and the provided algorithm. Eq. (14) states that the OT distance is evaluated against Dt \ D_fold (the held-out portion), while Algorithm 2, line 11 compares dOT(D_s_fold ∪ D_k, D_t_fold), i.e., against the selection fold D_fold. Additionally, the text says '1/k of Dt is used for selection, while the OT distance is evaluated against the remaining (k−1)/k data,' but Algorithm 2 sets D_tar ← Dt \ D_fold and uses D_tar for the kNN candidate search, while line 11 evaluates against D_fold. These two readings are opposite. Please align the text and pseudocode and state which protocol was actually implemented, since the reported OTM selection ratios and downstream gains depend on this choice.
  3. [§3.3–§3.4 and §4.2/§4.4] The comparison is not controlled for data weighting. In the segmentation experiments (Section 4.2), TAROT applies OT-potential weights with R = N + M while baselines use uniform repetition; in the instruction-tuning experiments (Section 4.4), TAROT-OTM is weighted with R = 0.5%N. Thus the gains of TAROT over baselines in those sections may stem from the weighting scheme rather than from the selection rule. The motion-prediction experiments (Section 4.3) do not apply weighting and still show gains, which partially mitigates this concern, but the paper should either apply the same repetition/weighting protocol to all methods or report the contribution of weighting explicitly.
  4. [§3.3, Algorithm 1] The fixed-size selection scheme is described as minimizing the OT distance, but Algorithm 1 is in fact a greedy k-nearest-neighbor procedure: it adds the k-th nearest candidates for each target point, and only in the final partial iteration does it use the OT dual potential (line 9). This is a heuristic approximation, not an empirical OT minimizer, and the paper does not characterize how close the resulting subset is to the OT-optimal subset. Given that the paper's title and framing emphasize optimal transport, please either provide a theoretical or empirical justification for this kNN greedy approximation, or soften the claim that the selected subset minimizes the OT distance.
minor comments (5)
  1. [§2.1, Eq. (2)] Eq. (2) uses N as the summation limit for checkpoint indexing, but N is also used for the candidate dataset size earlier in the paper; Eq. (4) correctly uses T for the number of checkpoints. Please make the notation consistent.
  2. [Algorithm 1, line 3] The input line states 'Desired subset size S = N', but the algorithm is intended for S ≤ N. This is presumably a typo; please correct it.
  3. [§3.2, Eqs. (6)–(8)] The whitening procedure does not specify over which data the centered gradients and covariance matrix are computed: the candidate pool Dc, the target set Dt, or their union. This affects the interpretation of the whitened distances and should be stated explicitly.
  4. [Table 1 and §4.4] The text says TAROT 'consistently outperforms all baseline methods,' but Table 1 shows TAROT-5% is slightly below LESS-5% on Qwen-2.5-7B MMLU (74.1 vs. 74.3) and TAROT-OTM matches LESS-5% on LLaMA-3.1-8B MMLU (65.7 vs. 65.7). Please qualify the 'consistently outperforms' claim accordingly.
  5. [§4.2–§4.4] The experimental sections do not report standard deviations or multiple seeds. Given the small reported improvements in some comparisons (e.g., less than 1 point on MMLU), adding variance estimates or at least stating the number of runs would strengthen the claims.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: TAROT's selection pipeline is a forward procedure validated on held-out benchmarks.

full rationale

The paper's derivation chain is not circular. WFD features are computed from gradients of a pretrained model (Eqs. 4-10), the subset is chosen by a greedy nearest-neighbor algorithm that approximately minimizes the OT distance to the target distribution (Eqs. 11-13, Algorithms 1-2), and the selected subset is then used to train a separate model that is evaluated on held-out validation sets. No parameter is fitted to the target performance, and the OTM selection ratio is determined by an OT-distance stopping criterion (Eq. 14), not by downstream accuracy. The method is therefore a genuine forward selection-then-train pipeline. The comparison to baselines changes both the distance metric and the selection rule, so the individual contribution of the OT formulation is not fully isolated; the main text also contains an internal inconsistency between Eq. (14), which evaluates OT distance against Dt \ D_fold, and Algorithm 2, which uses D_fold in the stopping rule. These are attribution and correctness concerns, not circularity. The stated limitation about overfitting on small target datasets is an honest caveat, and the self-citations (e.g., Feng et al. 2024 for the UniTraj framework) are infrastructural and not load-bearing. The central empirical claims are checked against external baselines and held-out evaluation, so no circular step can be exhibited.

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

The method introduces no new physical entities or model components beyond the selection algorithm. The free parameters are standard ML hyperparameters (checkpoints, projection dimension, k, R) that are not fitted to the final evaluation metric but are manually chosen. The axioms are standard domain assumptions for data selection methods.

free parameters (4)
  • Number of checkpoints T = 4 (segmentation), 1 (motion), unspecified for LLM
    The gradient feature sums over T checkpoints (Eq. 4). The choice of T is a hyperparameter affecting the feature quality; the paper reports using the last four checkpoints for segmentation but does not tune or analyze T systematically.
  • Random projection dimension (from Park et al.) = Not specified in paper
    The random projection in Eq. (6) is adopted from TRAK; the projection dimension is a hyperparameter that affects the whitened covariance estimation, but the paper does not report its value or sensitivity.
  • Number of folds k in OTM = 10
    Section 3.3: 'In our experiments, k = 10 ensures a good match with the target distribution'. Figure 8 shows the choice of k affects performance, so k is a tuned hyperparameter.
  • Repetition factor R in data weighting = N + M (segmentation), 0.5%N (instruction tuning), not used (motion)
    Section 3.4 defines R, and each experiment sets R differently. It controls the training steps and is not derived from first principles.
assumptions (4)
  • domain assumption Gradient features from a model trained on the candidate pool capture task-relevant data influence (Eq. 4).
    The entire method rests on the assumption that the sum of parameter gradients over checkpoints is a meaningful representation for selecting data for the target task. This is inherited from TRAK/LESS and is plausible but not proven, especially under domain shift between candidate and target distributions.
  • domain assumption The optimal transport distance between whitened gradient features is monotonic with downstream model generalization (Eq. 11).
    The paper justifies this via reference to Just et al. (2023) and empirical results, but it is a modeling assumption: minimizing feature-space OT is not proven to minimize task loss.
  • domain assumption Gradient features are sparse in high dimensions, making nearest-neighbor greedy selection a valid approximation of OT minimization (Section 3.3).
    The paper states this as a property ('most of the transport mass in the OT plan is concentrated on the nearest neighbors'), but no formal sparsity bound or error analysis is given.
  • domain assumption The target dataset Dt is a finite sample that faithfully represents the target distribution.
    OT is computed on finite samples; the paper uses validation splits as targets. If the target dataset is small or biased, selection will overfit to it (the paper's own limitation statement concedes this).

how reviews work

0 comments
Cite this review

Pith. "Pith review of TAROT: Targeted Data Selection via Optimal Transport." pith.science (2026). https://pith.science/paper/EDOGVJXQ

@misc{pith2026241200420,
  author       = {Pith},
  title        = {Pith review of: TAROT: Targeted Data Selection via Optimal Transport},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EDOGVJXQ}},
  note         = {Machine review of arXiv:2412.00420}
}
read the original abstract

We propose TAROT, a targeted data selection framework grounded in optimal transport theory. Previous targeted data selection methods primarily rely on influence-based greedy heuristics to enhance domain-specific performance. While effective on limited, unimodal data (i.e., data following a single pattern), these methods struggle as target data complexity increases. Specifically, in multimodal distributions, these heuristics fail to account for multiple inherent patterns, leading to suboptimal data selection. This work identifies two primary factors contributing to this limitation: (i) the disproportionate impact of dominant feature components in high-dimensional influence estimation, and (ii) the restrictive linear additive assumptions inherent in greedy selection strategies. To address these challenges, TAROT incorporates whitened feature distance to mitigate dominant feature bias, providing a more reliable measure of data influence. Building on this, TAROT uses whitened feature distance to quantify and minimize the optimal transport distance between the selected data and target domains. Notably, this minimization also facilitates the estimation of optimal selection ratios. We evaluate TAROT across multiple tasks, including semantic segmentation, motion prediction, and instruction tuning. Results consistently show that TAROT outperforms state-of-the-art methods, highlighting its versatility across various deep learning tasks. Code is available at https://github.com/vita-epfl/TAROT.

Figures

Figures reproduced from arXiv: 2412.00420 by the authors.

Figure 1
Figure 1. Left: the selection objective of TAROT, designed to minimize optimal transport costs. The method operates on a target distribution and a candidate data pool, selecting a subset from the candidate data to reduce transport costs. Right: the approximate performance improvements and data efficiency achieved by TAROT across various tasks. It illustrates that TAROT consistently enhances task per￾formance in different doma… view at source ↗
Figure 2
Figure 2. Overview of TAROT. Left: Raw gradient features are whitened and projected onto a unit hypersphere. Right: A greedy algorithm iteratively selects candidate samples. During the second iteration, adding two new distant data points increases the overall transport cost, prompting the algorithm to stop. As a result, only in-distribution data points are selected to minimize the OT distance. as a single task (e.g. code gene… view at source ↗
Figure 3
Figure 3. Illustration of target examples and their corresponding closest (most helpful) and furthest (most detracting) [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: WFD outperforms baseline methods on both tasks. The x-axis indicates the number of model checkpoints that a given method uses to compute attribution scores. The y-axis indicates the method’s efficacy as measured by LDS. values. Additionally, we employ ZCA whitening ins…
Figure 7
Figure 7. Figure 7: Results of Targeted Selection for Motion Predic￾tion. The figure illustrates the performance of Wayformer under various data selection methods. Main Results [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]
Figure 6
Figure 6. Figure 6: Data selection visualization in gradient feature space with T-SNE. We select 10% of Cityscapes’ training set, with validation split as the target. Left: Samples se￾lected by DsDm cluster tightly around the target distribution center. Right: Samples selected by TAROT ar…
Figure 8
Figure 8. Figure 8: Ablation study for the choice of k. The figure shows the performance of Wayformer under different k. Main Results [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 9
Figure 9. Figure 9: More Qualitative Results of WFD in Motion Prediction. Visualization of motion prediction tasks using the nuScenes and Waymo datasets. The dark-to-light blue lines indicate the past (input) and future (ground truth) trajectories for the target vehicle, while green lines…
Figure 10
Figure 10. Figure 10: More Qualitative Results of WFD in Semantic Segmentation. Semantic segmentation tasks with examples from the Cityscapes and GTA5 datasets. For each target example, the top image shows the RGB scene, and the bottom image depicts the corresponding ground truth semantic …

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. GORACS: Group-level Optimal Transport-guided Coreset Selection for LLM-based Recommender Systems

    cs.IR 2025-06 conditional novelty 6.0 of 10

    GORACS selects small groups of fine-tuning examples via an optimal-transport and gradient-norm proxy objective, outperforming prior coreset methods for LLM-based recommendation.

Reference graph

Works this paper leans on

20 extracted references · 6 canonical work pages · cited by 1 Pith paper

  1. [1]

    L., Almeida, D., Altenschmidt, J., Altman, S., Anadkat, S., et al

    Achiam, J., Adler, S., Agarwal, S., Ahmad, L., Akkaya, I., Aleman, F. L., Almeida, D., Altenschmidt, J., Altman, S., Anadkat, S., et al. Gpt-4 technical report. arXiv preprint arXiv:2303.08774,

    arXiv

  2. [5]

    Details for Algorithms We present the detailed data selection algorithm described in Section 3.3 in Algorithms 1 and 2 D

    C. Details for Algorithms We present the detailed data selection algorithm described in Section 3.3 in Algorithms 1 and 2 D. NuScenes Leaderboard Performance In the motion prediction task, TAROT helps Wayformer achieve the first place on the nuScenes leaderboard (Ta- ble 7). surpassing MTR-UniTraj (Feng et al., 2024), which relies on large-scale datasets ...

    work page 2024

  3. [7]

    Dsdm: Model- aware dataset selection with datamodels

    Engstrom, L., Feldmann, A., and Madry, A. Dsdm: Model- aware dataset selection with datamodels. arXiv preprint arXiv:2401.12926,

    arXiv

  4. [9]

    A., Kahou, S

    9 TAROT: Targeted Data Selection via Optimal Transport Girgis, R., Golemo, F., Codevilla, F., Weiss, M., D’Souza, J. A., Kahou, S. E., Heide, F., and Pal, C. Latent variable sequential set transformers for joint multi-agent motion prediction. arXiv preprint arXiv:2104.00563,

    arXiv

  5. [11]

    Tsds: Data selection for task-specific model finetuning

    Liu, Z., Karbasi, A., and Rekatsinas, T. Tsds: Data selection for task-specific model finetuning. In The Thirty-eighth Annual Conference on Neural Information Processing Systems. Liu, Z., Zhou, K., Zhao, W. X., Gao, D., Li, Y ., and Wen, J.-R. Less is more: Data value estimation for visual in- struction tuning. arXiv preprint arXiv:2403.09559,

    arXiv

  6. [12]

    S., and Sapp, B

    Nayakanti, N., Al-Rfou, R., Zhou, A., Goel, K., Refaat, K. S., and Sapp, B. Wayformer: Motion forecasting via simple & efficient attention networks. In 2023 IEEE International Conference on Robotics and Automation (ICRA), pp. 2980–2987. IEEE,

    2023

  7. [13]

    Dinov2: Learning robust visual features without supervision

    Oquab, M., Darcet, T., Moutakanni, T., V o, H., Szafraniec, M., Khalidov, V ., Fernandez, P., Haziza, D., Massa, F., El- Nouby, A., et al. Dinov2: Learning robust visual features without supervision. arXiv preprint arXiv:2304.07193,

    arXiv

  8. [14]

    R., Vineet, V ., Roth, S., and Koltun, V

    Richter, S. R., Vineet, V ., Roth, S., and Koltun, V . Play- ing for data: Ground truth from computer games. In Computer Vision–ECCV 2016: 14th European Confer- ence, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part II 14, pp. 102–118. Springer,

    work page 2016

Show all 20 references
  1. [15]

    Semantic- former: Holistic and semantic traffic scene representation for trajectory prediction using knowledge graphs

    Sun, Z., Wang, Z., Halilaj, L., and Luettin, J. Semantic- former: Holistic and semantic traffic scene representation for trajectory prediction using knowledge graphs. arXiv preprint arXiv:2404.19379,

    arXiv

  2. [16]

    W., Chowdhery, A., Le, Q., Chi, E., Zhou, D., et al

    10 TAROT: Targeted Data Selection via Optimal Transport Suzgun, M., Scales, N., Sch¨arli, N., Gehrmann, S., Tay, Y ., Chung, H. W., Chowdhery, A., Le, Q., Chi, E., Zhou, D., et al. Challenging big-bench tasks and whether chain-of- thought can solve them. InFindings of the Asso...

    2023

  3. [17]

    doi: 10.1609/aaai.v38i8. 28771. URL https://ojs.aaai.org/index. php/AAAI/article/view/28771. Yao, Z., Li, X., Lang, B., and Chuah, M. C. Goal-lbp: Goal-based local behavior guided trajectory prediction for autonomous driving. IEEE Transactions on Intelligent Transportation Systems,

    doi

  4. [18]

    More Related Work Data Influence Estimation

    11 TAROT: Targeted Data Selection via Optimal Transport A. More Related Work Data Influence Estimation. Pruthi et al. (2020) proposed a method to estimate data influence by tracing the gradi- ent descent process. Another approach utilizes datamodels: Ilyas et al. (2022) showed...

    2020

  5. [19]

    Kang et al

    and sensitivity analysis in data val- uation (Just et al., 2023). Kang et al. (2024) applied it to optimize data selection ratios across multiple sources. We integrate influence estimation with optimal transport, enabling effective subset selection to boost target domain perfo...

    2024

  6. [2013]

    Efficient en- sembles improve training data attribution

    Deng, J., Li, T.-W., Zhang, S., and Ma, J. Efficient en- sembles improve training data attribution. arXiv preprint arXiv:2405.17293,

    arXiv

  7. [2016]

    Aligning ai with shared human values

    Hendrycks, D., Burns, C., Basart, S., Critch, A., Li, J., Song, D., and Steinhardt, J. Aligning ai with shared human values. Proceedings of the International Conference on Learning Representations (ICLR), 2021a. Hendrycks, D., Burns, C., Basart, S., Zou, A., Mazeika, M., Song,...

    arXiv

  8. [2017]

    K., Ahn, H., Bae, J., Zhao, K., Kang, M., Chung, Y ., Pratapa, A., Neiswanger, W., Strubell, E., Mitamura, T., et al

    Choe, S. K., Ahn, H., Bae, J., Zhao, K., Kang, M., Chung, Y ., Pratapa, A., Neiswanger, W., Strubell, E., Mitamura, T., et al. What is your data worth to gpt? llm-scale data valuation with influence functions. arXiv preprint arXiv:2405.13954,

    arXiv

  9. [2020]

    Rethinking atrous convolution for semantic image segmentation

    Chen, L.-C. Rethinking atrous convolution for semantic image segmentation. arXiv preprint arXiv:1706.05587,

    arXiv

  10. [2021]

    Feng, L., Bahari, M., Amor, K. M. B., Zablocki, ´E., Cord, M., and Alahi, A. Unitraj: A unified framework for scalable vehicle trajectory prediction. arXiv preprint arXiv:2403.15098,

    arXiv

  11. [2023]

    M., Longpre, S., Lambert, N., Wang, X., Muennighoff, N., Hou, B., Pan, L., Jeong, H., et al

    Albalak, A., Elazar, Y ., Xie, S. M., Longpre, S., Lambert, N., Wang, X., Muennighoff, N., Hou, B., Pan, L., Jeong, H., et al. A survey on data selection for language models. arXiv preprint arXiv:2402.16827,

    arXiv

  12. [2024]

    The llama 3 herd of models

    Dubey, A., Jauhri, A., Pandey, A., Kadian, A., Al-Dahle, A., Letman, A., Mathur, A., Schelten, A., Yang, A., Fan, A., et al. The llama 3 herd of models. arXiv preprint arXiv:2407.21783,

    arXiv

Pith tools

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