Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

Attamba: Attending To Multi-Token States

T0 review · 3 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

Pith's one-line read State-space models can replace transformer key-value projections, compressing token chunks and improving perplexity at a matched memory footprint.

desk verdict A genuinely novel way to compress KV states, undone in the abstract by a baseline that is not actually footprint-matched. read the letter →

arxiv 2411.17685 v1 pith:4QPZWGJL submitted 2024-11-26 cs.LG cs.CL

classification cs.LGcs.CL
keywords state-spacemodelsattentionKV-cachecompressiontokenchunkingtransformerefficiencylanguagemodelingsequence
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 Attamba, a transformer variant that replaces the key and value projection matrices with state-space models, each of which compresses a chunk of P consecutive tokens into a single key or value state. Attention then runs over these compressed states rather than over every token, reducing both KV-cache memory and attention computation by roughly a factor of P. The authors report that on language modeling, Attamba improves perplexity by 24% relative to a transformer matched for KV-cache and attention footprint, and can trade about 5% perplexity for a roughly 4x smaller KV-cache and attention FLOP count. The broader claim is that SSM compression is not a lossy shortcut but a workable replacement for token-level attention, with flexible chunk boundaries that allow a smooth transition between quadratic and linear scaling.

What carries the argument

The KV-SSM block is the load-bearing component: a selective state-space model that reads a chunk of P tokens and outputs a single hidden state per chunk, replacing both the $W_K$ and $W_V$ projections. The chunk-aware causal mask $M_{\text{train}}$ (Equation 5) then makes queries attend to the last token of each completed chunk and to tokens in the current partial chunk, while $M_{\text{test}}$ (Equation 7) restricts inference attention to chunk boundaries; when a sliding window of L leading tokens is kept uncompressed (Equation 8), the mask combines boundary states with the most recent tokens. Cyclic chunking shifts the chunk offset by layer index to spread boundary bias across layers. Together these pieces turn the attention map from $n \times n$ into $(n/P + L) \times (n/P + L)$, which is where both the memory and FLOP savings come from.

What would settle it

Measure peak GPU memory and attention FLOPs for Attamba and for the iso-KV-plus-sliding-window transformer on a long sequence (say 4096 tokens) at the same batch size, and retrain or evaluate both on a long-context benchmark; if the transformer matches Attamba's perplexity once the two are compared at equal measured footprint, the core efficiency-advantage claim collapses.

Watch

Extended reading notes

Core claim

The paper's central claim is that a state-space model can serve as the key and value projector of a transformer, compressing every chunk of P tokens into one state that attention then treats as a single token. Formally, the key and value sequences are replaced by $K^{(p)} = \mathrm{SSM}_K(X^{(p)})$ and $V^{(p)} = \mathrm{SSM}_V(X^{(p)})$ for each chunk, and a causal mask restricts attention to the final state of completed chunks plus the current partial chunk. At inference only the boundary states are cached, so both KV-cache size and the attention map shrink by a factor of P. The authors report that this compressed attention is not simply a cheaper approximation: on WikiText-2, Attamba reaches 21.74 perplexity in an 8-billion-token training run, beating an iso-KV transformer (22.89) and an iso-KV-plus-sliding-window transformer (28.66), and the abstract's headline comparison states a 24% perplexity improvement over a transformer of similar KV-cache and attention footprint. They also show the SSM compresses robustly across chunking strategies, including randomized boundaries, which they take as evidence that attention over compressed states is a flexible alternative to token-level attention.

Load-bearing premise

The footprint-matched comparison assumes that setting a transformer's attention dimension to $F = E/P$ and evaluating with a sliding window of $L/P$ tokens reproduces Attamba's KV-cache and attention-map cost; if those baselines are not truly equal in measured memory and FLOPs, the claimed 24% perplexity advantage is not established.

Editorial extensions

If this is right

  • At inference, only the final SSM state of each chunk is cached, so KV-cache size and attention-map size shrink by a factor of P; with chunk size 8 and 64 leading tokens, the paper reports roughly 8x KV-cache compression at a 10% perplexity cost.
  • Variable-length chunk boundaries, implemented via cu_seqlens, let the same model move between quadratic and linear attention by adjusting how many tokens each SSM state compresses.
  • Cyclic chunking (shifting chunk boundaries by layer index) improves perplexity by about 5% over fixed boundaries, and randomized boundaries match uniform chunking, indicating the compression is not tied to a specific token grouping.
  • Removing the explicit key and value projection matrices costs only about 1% perplexity, so the SSM block can serve as a drop-in replacement for the projection weights.

Reading between the lines

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

  • The paper evaluates only WikiText-2, a highly local task; a more demanding test would be long-context retrieval or needle-in-a-haystack, where compressed chunk states must preserve distal information rather than recent locality.
  • Robustness to randomized boundaries suggests a learned or importance-based boundary predictor could trade a small perplexity penalty for large compute savings at inference, extending the paper's fixed/cyclic chunking to adaptive chunking.
  • The claimed 24% gain rests on derived iso-KV and iso-FLOP baseline formulas (Appendix A); directly measuring peak memory and wall-clock time of Attamba against those baselines on the same hardware would make the efficiency comparison robust.
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

3 major / 4 minor

Summary. The paper proposes Attamba, a transformer variant in which the key and value projections are replaced by SSM blocks that compress chunks of P tokens into single states, followed by attention over these compressed states plus a set of leading (uncompressed) tokens. The authors report WikiText2 perplexity for roughly 60M-parameter models trained on 1B or 8B tokens and claim a 24% perplexity improvement over a transformer with similar KV-cache and attention footprint, as well as a roughly 4x reduction in KV-cache and attention FLOPs for a 5% perplexity trade-off. The manuscript also studies chunk-boundary strategies, SSM state dimension, leading-token count, and pseudo-chunking, with code and training logs provided.

Significance. If the architecture is causally sound and the efficiency-matched baselines are genuine, Attamba is a useful hybrid that combines SSM-based local compression with global attention over compressed states, offering a controllable quality/efficiency knob and a path from quadratic to linear attention. The paper includes several strengths: the efficiency arithmetic is straightforward, the design choices are ablated, and the authors provide logs and code, which supports reproducibility. However, the central efficiency claim currently rests on transformer baselines whose attention footprint is much smaller than Attamba's, and the test-time masking rule appears to permit attention to future tokens within the current chunk; both issues must be resolved before the headline claims can be accepted.

major comments (3)
  1. [§5.1, Figure 7, Appendix A Eqs. (9)-(11)] The 'similar footprint' transformer baseline is not footprint-matched. The +KVC+SWA baseline uses attention dimension F = E/P = 128 and a sliding window of size L_lead/P = 8, whereas Attamba with P=4 and L_lead=32 attends to roughly n/P + L_lead = 288 key-value states of dimension E = 512. For n = 1024 the baseline attention map is therefore about 36x smaller, and its attention FLOPs are even smaller because the dimension is also reduced. The reported 24% perplexity gap (21.74 vs. 28.66 in Figure 6/20) may thus reflect the baseline's severe restriction to eight recent tokens rather than a benefit of SSM compression. Please construct a transformer baseline that matches Attamba's full cache, including leading tokens, and its attention map size, e.g., F satisfying 2BLF = 2B(n/P + L_lead)E and window W = n/P + L_lead, or clearly report the residual mismatch.
  2. [§4.1, Eq. (7)] The test-time mask Mtest appears to violate causality. For a query at position i inside chunk p = floor(i/P), the condition j <= floor(i/P) includes the compressed state K(p)[-1], which is the SSM output after processing the entire chunk, including tokens after i. At training time Eq. (5) avoids this by allowing only partial-chunk outputs for the current chunk, but at test time those partial outputs are discarded. If Attamba is evaluated with the stated mask and cached chunk-boundary states, the model can attend to future tokens within the current chunk. Please either restrict test-time attention to completed chunks only (with the current chunk handled by leading tokens), or maintain causally valid partial states; the current formulation needs clarification or correction.
  3. [§5, §5.2, Appendix B] The training budget is stated inconsistently. Section 5 says 'Unless otherwise specified, we train on approximately 1B tokens', Appendix B repeats the 1B budget, but Section 5.2 and Figures 6/20 report training 'for 100,000 steps over 8 billion tokens'. Since the headline comparison with Mamba, minGRU, Hawk, and the Transformer baselines comes from the 8B run, the paper needs to state which budget applies to each figure and why the appendix describes a different setting; otherwise the results are not reproducible.
minor comments (4)
  1. [Notation throughout] The symbol L is used both for sequence length (Appendix A, Eqs. (9)-(11)) and for the number of leading tokens (main text, Figure 5, Figure 7 caption). This ambiguity makes the baseline construction in Eq. (9) and the caption 'window size = L/P' difficult to interpret; please use distinct symbols, e.g., N_seq and L_lead.
  2. [Appendix A.1.1, A.2] There are typographical errors: 'intorduce' should be 'introduce', 'Psuedo-chunking' should be 'Pseudo-chunking', and 'inStuffed Mamba' in Related Work is missing a space.
  3. [§4.1, Figure 13] The 'Attamba-Linear' variant and the claimed smooth transition between quadratic and linear scaling are motivated but not evaluated; a small experiment or explicit statement that this is future work would make the claim more precise.
  4. [§6, Figure 17] The limitations section is candid about the small-scale, local-task evaluation, and Figure 17 correctly shows that a true 128x KV-cache reduction with only one leading token performs poorly. These statements appropriately qualify the generality of the results.

Circularity Check

1 steps flagged · score 2.0 of 10

No derivation-level circularity; the central quality claim is empirical. The only mild circularity is that the 'footprint-matched' transformer baselines are defined from Attamba's own chunking hyperparameters, making the comparison partly self-defined rather than an independent control.

  1. self definitional [Section 5.1 and Figure 7 caption; Appendix A, Eq. 9]
    "Specifically, for transformers, we emulate smaller KV-Cache sizes by reducing the attention model dimension F such that F = E/P, and smaller attention maps by employing sliding window attention (SWA) during evaluation. ... to match the attention map size, we evaluate these models in Sliding Window attention with window size = L/P."

    The 'matched' transformer baseline's footprint is defined directly from Attamba's own hyperparameters (F = E/P and window = L/P), so the 'similar KV-Cache and attention footprint' claim is partly fixed by definition rather than independently established. Attamba's actual test-time cache and attention map include the leading-token block, giving roughly n/P + L keys (Eq. 8), while the constructed sliding-window baseline uses only L/P keys and a reduced attention dimension F. Thus the comparison baseline is constructed to be smaller than Attamba, and the headline footprint match is a definitional artifact rather than a measured equality. The perplexity advantage itself is still an empirical result, so this is a mild baseline-construction circularity, not a derivation-level one.

full rationale

The paper does not derive its quality claim from an equation that reduces to a fitted constant. Attamba is an architecture proposal evaluated empirically: models are trained on a dclm-baseline subset and evaluated on WikiText2 test perplexity, with pseudo-chunking (Figure 16) and ablations (Figures 11, 12, 14, 15) providing independent internal evidence. The self-citations present in the paper (Akhauri et al., 2024 for token importance; Palu, which shares an author) are related-work or motivational statements, not load-bearing premises: the token-importance self-citation motivates chunk-boundary selection, but the paper's own experiments show FAttn and FSSM chunking do not help, so the final method does not depend on that cited result. There is no imported uniqueness theorem and no ansatz smuggled in solely via self-citation. The only mild circularity is the construction of 'fair' transformer baselines using Attamba's own P and L (F = E/P, sliding-window size = L/P), which makes the footprint match definitional and arguably understates the baseline's attention restriction. This is an experimental-design fairness concern rather than a case where a prediction reduces to its inputs by construction, so the circularity score is low.

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

The central claims rest on a small set of hand-chosen hyperparameters (P, L, Ds, cyclic offset), on the domain assumption that SSM compression preserves attention-relevant information, and on the ad hoc comparison methodology for iso-footprint baselines. No new physical or mathematical entities are introduced.

free parameters (4)
  • Chunk size P = 4, 8, 64, 128 (default 4 or 8)
    Chosen by hand; controls the compression factor and the quality/efficiency trade-off. The paper reports perplexity for several values and defaults to P=4 or P=8.
  • Leading tokens L = 32 (for P=4), 8, 16, 64, 128, 1024 in ablations
    Number of recent tokens that retain uncompressed attention; tuned per chunk size and used to construct the 'similar footprint' baselines.
  • SSM state dimension Ds = 32 (default)
    Chosen after an ablation (Figure 12) showing diminishing returns beyond Ds=32; affects parameter overhead and representation capacity.
  • Cyclic chunking layer offset = layer index
    Shifts chunk boundaries per layer to reduce fixed-boundary bias; a design choice validated in Figure 14.
assumptions (4)
  • domain assumption Selective state-space models (Mamba) can compress a chunk of tokens into a single state that retains information useful for attention (Section 3.2, Eq. 3).
    The method relies on SSM compression preserving enough information for downstream attention; this is plausible but tested only on WikiText2.
  • domain assumption WikiText2 test perplexity is a valid proxy for language-model quality for the claims made (Section 5).
    All conclusions are drawn from this single, highly local benchmark; long-context behavior is explicitly untested.
  • ad hoc to paper The test-time attention mask (Mtest in Eq. 7) enforces causality when attending to compressed chunk states.
    As written, j <= floor(i/P) lets a query in the middle of chunk c attend to the final state of chunk c, which aggregates future tokens; the paper does not explain how the final state of an incomplete chunk is available causally.
  • ad hoc to paper The iso-KV/iso-FLOPs baseline formulas (Eqs. 9-11) capture the dominant costs of attention and KV-cache, making the reduced-dimension transformers fair comparisons.
    The equations drop normalization, softmax, and activation costs, and the sliding-window baseline's attention map is much smaller than Attamba's, so the 'similar footprint' premise is questionable.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Attamba: Attending To Multi-Token States." pith.science (2026). https://pith.science/paper/4QPZWGJL

@misc{pith2026241117685,
  author       = {Pith},
  title        = {Pith review of: Attamba: Attending To Multi-Token States},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4QPZWGJL}},
  note         = {Machine review of arXiv:2411.17685}
}
read the original abstract

When predicting the next token in a sequence, vanilla transformers compute attention over all previous tokens, resulting in quadratic scaling of compute with sequence length. State-space models compress the entire sequence of tokens into a fixed-dimensional representation to improve efficiency, while other architectures achieve sub-quadratic complexity via low-rank projections or sparse attention patterns over the sequence. In this paper, we introduce Attamba, a novel architecture that uses state-space models to compress chunks of tokens and applies attention on these compressed key-value representations. We find that replacing key and value projections in a transformer with SSMs can improve model quality and enable flexible token chunking, resulting in 24% improved perplexity with transformer of similar KV-Cache and attention footprint, and ~4 times smaller KV-Cache and Attention FLOPs for 5% perplexity trade-off. Attamba can perform attention on chunked-sequences of variable length, enabling a smooth transition between quadratic and linear scaling, offering adaptable efficiency gains.

Figures

Figures reproduced from arXiv: 2411.17685 by the authors.

Figure 1
Figure 1. Attamba uses State-Space Models (SSM) to compress key-value sequences into token chunks (e.g., chunks of P = 4), reducing the attention map and KV-Cache size by P× by storing only chunk boundaries. 1 arXiv:2411.17685v1 [cs.LG] 26 Nov 2024 [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. State Space Models (SSMs) efficiently encode multiple tokens into a single representation. By compressing key (K) and value (V ) sequences into chunked representations, SSMs maintain essential contextual information, enabling efficient query (Q) inter￾actions. This approach minimizes KV-Cache size by storing only chunk boundaries and reduces the computational cost of attention. Attamba demonstrates robustness to ran… view at source ↗
Figure 3
Figure 3. Attamba uses SSM blocks to compress chunks of tokens (P = 4 in the example above) into a single token. 3 [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (13 more)
Figure 4
Figure 4. Figure 4: Full-Attention has a purely causal mask, attending to all past tokens. Attamba uses Key-Value SSM blocks to compress chunks of P tokens (e.g. P = 4) into one state. Tokens compressed by SSMs are at chunk boundaries. This is incorporated with a sliding-window attention …
Figure 5
Figure 5. Figure 5: Leading-Tokens (L) control how many ’leading’ tokens full-attention happens over, preserving full-attention on the newest tokens. This resembles Sliding-Window attention. Chunk-size (P) controls how many tokens are chunked by the SSM. Other Design Considerations: In de…
Figure 6
Figure 6. Figure 6: Comparing Attamba with SSMs (Mamba), minGRU, Hawk and Transformers (Xmer) by training on 8 billion tokens. E, P, L, Ds, G, H denote Model-Dim, Chunk Size, Leading Tokens, SSM State-Dim, Num. Groups and Num. Heads respectively, 0 when not applicable. Models ∈ [60, 64]M …
Figure 7
Figure 7. Figure 7: Comparing Attamba with a base transformer with match￾ing parameter counts. Further, we train variants with smaller KV￾Cache size to match Attamba. Additionally, to match the attention map size, we evaluate these models in Sliding Window attention with window size = L P…
Figure 8
Figure 8. Figure 8: Each head in Llama-2-7B attends to tokens in a manner that is largely uncorrelated (Kendall-Tau ∈ [−0.2, 0.8]) with other heads. A. Appendix A.1. On Token Chunking Processing sequences in fixed-size chunks simplifies im￾plementation, but can limit models flexibility. P…
Figure 9
Figure 9. Figure 9: Different token-chunking strategies we investigate. L, T, C represent layer, token and chunk respectively. Value Aggregation Attention KSSM KSSM KSSM KSSM Test-Time Train-Time Train-Time Attention Mask VSSM VSSM VSSM VSSM VSSM Attamba-Nsq KSSM Test-Time Attention Mask …
Figure 10
Figure 10. Figure 10: Attamba employs Key and Value State-Space Models (SSMs) to accumulate local information within chunks of tokens. At test time, only the final accumulated activations from each chunk are used in the standard attention mechanism. The red lines denote the auto-regressive…
Figure 11
Figure 11. Figure 11: Removing the Key-Value projection matrices when us￾ing K-V SSMs does not impact WikiText2 test-perplexity signifi￾cantly. 40000 50000 60000 Global Step 34 36 38 40 42 44 WK2 Perplexity Impact of Increasing SSM State-Dimension Attamba Ds 32 (62M) 34.6 Attamba Ds 128 (6…
Figure 13
Figure 13. Figure 13: Attamba-Linear maintains linear complexity, by having a fixed-size attention, and dividing the sequence length (L) into chunks. Attamba-Quadratic has quadratic complexity (albeit lower FLOPs/Memory than standard transformer) as the SSM only processes P tokens. w, r, g…
Figure 15
Figure 15. Figure 15: Chunk size of 128 implies a 128× smaller KV-Cache. It outperforms Chunk 4/8/64 because we do full-attention on partial-chunks, giving significant advantage as chunk-size in￾creases on local evaluation tasks like WikiText2. cantly benefit from research in token importa…
Figure 16
Figure 16. Figure 16: Pseudo-Chunking (replacing Key-Value projection ma￾trices with SSMs, but attending to all tokens) can marginally improve transformer perplexity. (C: Chunk Size) 10 0 10 1 10 2 L (Leading Uncompressed Tokens) 30 35 40 45 50 55 60 65 WK2 Perplexity Impact of Full-Attent…
Figure 18
Figure 18. Figure 18: iso-Parameter and iso-FLOPs still has higher memory overhead and does not address the L 2 attention and KV-Cache overhead. 0 20 40 60 80 100 120 P (Chunk Size) 0 20 40 60 80 100 120 Model Dimension Ratio Model Dimension Ratio vs. Chunk Size (P) Iso-FLOP Iso-KV-Cache I…
Figure 20
Figure 20. Figure 20: Comparing Attamba with SSMs (Mamba), minGRU, Hawk and Transformers (Xmer) by training on 8 billion tokens. E, C, L, Ds, G, H denote Model-Dim, Chunk-Size, Leading-Tokens, SSM State-Dim, Num. Groups and Num. Heads respectively, 0 when not applicable. Models ∈ [60, 64]M…

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. DART: Decoded Attention over Recurrent States for Efficient Long-Context Sequence Modeling

    cs.LG 2026-08 conditional novelty 6.0 of 10

    DART augments Mamba-2 with attention over its own chunk state contributions, improving associative recall and retrieval with a smaller inference cache than token-level attention.

Reference graph

Works this paper leans on

24 extracted references · 10 canonical work pages · cited by 1 Pith paper

  1. [1]

    F., Dotzel, J., Zhang, Z., Rush, A

    Akhauri, Y., AbouElhamayed, A. F., Dotzel, J., Zhang, Z., Rush, A. M., Huda, S., and Abdelfattah, M. S. S hadow LLM : Predictor-based contextual sparsity for large language models. In Al-Onaizan, Y., Bansal, M., and Chen, Y.-N. (eds.), Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pp.\ 19154--19167, Miami, Florida...

  2. [2]

    S., and Wu, K.-C

    Chang, C.-C., Lin, W.-C., Lin, C.-Y., Chen, C.-Y., Hu, Y.-F., Wang, P.-S., Huang, N.-C., Ceze, L., Abdelfattah, M. S., and Wu, K.-C. Palu: Compressing kv-cache with low-rank projection, 2024. URL https://arxiv.org/abs/2407.21118

  3. [3]

    Stuffed mamba: State collapse and state capacity of rnn-based long-context modeling

    Chen, Y., Zhang, X., Hu, S., Han, X., Liu, Z., and Sun, M. Stuffed mamba: State collapse and state capacity of rnn-based long-context modeling. arXiv preprint arXiv:2410.07145, 2024

  4. [4]

    M., Likhosherstov, V., Dohan, D., Song, X., Gane, A., Sarlos, T., Hawkins, P., Davis, J

    Choromanski, K. M., Likhosherstov, V., Dohan, D., Song, X., Gane, A., Sarlos, T., Hawkins, P., Davis, J. Q., Mohiuddin, A., Kaiser, L., et al. Rethinking attention with performers. In International Conference on Learning Representations

  5. [5]

    and Gu, A

    Dao, T. and Gu, A. Transformers are ssms: Generalized models and efficient algorithms through structured state space duality. In Forty-first International Conference on Machine Learning

  6. [6]

    L., Fernando, A., Botev, A., Cristian-Muraru, G., Gu, A., Haroun, R., Berrada, L., Chen, Y., Srinivasan, S., et al

    De, S., Smith, S. L., Fernando, A., Botev, A., Cristian-Muraru, G., Gu, A., Haroun, R., Berrada, L., Chen, Y., Srinivasan, S., et al. Griffin: Mixing gated linear recurrences with local attention for efficient language models. arXiv preprint arXiv:2402.19427, 2024

  7. [7]

    O., Bengio, Y., and Hajimirsadegh, H

    Feng, L., Tung, F., Ahmed, M. O., Bengio, Y., and Hajimirsadegh, H. Were rnns all we needed?, 2024. URL https://arxiv.org/abs/2410.01201

  8. [8]

    Y., Roziere, B., Lopez-Paz, D., and Synnaeve, G

    Gloeckle, F., Idrissi, B. Y., Roziere, B., Lopez-Paz, D., and Synnaeve, G. Better & faster large language models via multi-token prediction. In Forty-first International Conference on Machine Learning

Show all 24 references
  1. [9]

    and Dao, T

    Gu, A. and Dao, T. Mamba: Linear-time sequence modeling with selective state spaces. arXiv preprint arXiv:2312.00752, 2023

  2. [10]

    Efficiently modeling long sequences with structured state spaces

    Gu, A., Goel, K., and Re, C. Efficiently modeling long sequences with structured state spaces. In International Conference on Learning Representations

  3. [11]

    Hippo: Recurrent memory with optimal polynomial projections

    Gu, A., Dao, T., Ermon, S., Rudra, A., and R \'e , C. Hippo: Recurrent memory with optimal polynomial projections. Advances in neural information processing systems, 33: 0 1474--1487, 2020

  4. [12]

    W., Jitsev, J., Kollar, T., Dimakis, A

    Li, J., Fang, A., Smyrnis, G., Ivgi, M., Jordan, M., Gadre, S., Bansal, H., Guha, E., Keh, S., Arora, K., Garg, S., Xin, R., Muennighoff, N., Heckel, R., Mercat, J., Chen, M., Gururangan, S., Wortsman, M., Albalak, A., Bitton, Y., Nezhurina, M., Abbas, A., Hsieh, C.-Y., Ghosh,...

  5. [13]

    Jamba: A hybrid transformer-mamba language model

    Lieber, O., Lenz, B., Bata, H., Cohen, G., Osin, J., Dalmedigos, I., Safahi, E., Meirom, S., Belinkov, Y., Shalev-Shwartz, S., et al. Jamba: A hybrid transformer-mamba language model. arXiv preprint arXiv:2403.19887, 2024

  6. [14]

    Pointer sentinel mixture models

    Merity, S., Xiong, C., Bradbury, J., and Socher, R. Pointer sentinel mixture models. In International Conference on Learning Representations, 2022

  7. [15]

    The illusion of state in state-space models

    Merrill, W., Petty, J., and Sabharwal, A. The illusion of state in state-space models. In Forty-first International Conference on Machine Learning

  8. [16]

    Samba: Simple hybrid state space models for efficient unlimited context language modeling

    Ren, L., Liu, Y., Lu, Y., Shen, Y., Liang, C., and Chen, W. Samba: Simple hybrid state space models for efficient unlimited context language modeling. arXiv preprint arXiv:2406.07522, 2024

  9. [17]

    Shadowkv: Kv cache in shadows for high-throughput long-context llm inference, 2024

    Sun, H., Chang, L.-W., Bao, W., Zheng, S., Zheng, N., Liu, X., Dong, H., Chi, Y., and Chen, B. Shadowkv: Kv cache in shadows for high-throughput long-context llm inference, 2024. URL https://arxiv.org/abs/2410.21465

  10. [18]

    Y., Haziza, D., Wehrstedt, L., Copet, J., Teytaud, O., and Lopez-Paz, D

    Videau, M., Idrissi, B. Y., Haziza, D., Wehrstedt, L., Copet, J., Teytaud, O., and Lopez-Paz, D. Meta lingua: A minimal PyTorch LLM training library, 2024. URL https://github.com/facebookresearch/lingua

  11. [19]

    and Li, Q

    Wang, S. and Li, Q. Stablessm: Alleviating the curse of memory in state-space models through stable reparameterization. In Forty-first International Conference on Machine Learning

  12. [20]

    Z., Khabsa, M., Fang, H., and Ma, H

    Wang, S., Li, B. Z., Khabsa, M., Fang, H., and Ma, H. Linformer: Self-attention with linear complexity. arXiv preprint arXiv:2006.04768, 2020

  13. [21]

    Efficient streaming language models with attention sinks

    Xiao, G., Tian, Y., Chen, B., Han, S., and Lewis, M. Efficient streaming language models with attention sinks. In The Twelfth International Conference on Learning Representations

  14. [22]

    A., Ainslie, J., Alberti, C., Ontanon, S., Pham, P., Ravula, A., Wang, Q., Yang, L., et al

    Zaheer, M., Guruganesh, G., Dubey, K. A., Ainslie, J., Alberti, C., Ontanon, S., Pham, P., Ravula, A., Wang, Q., Yang, L., et al. Big bird: Transformers for longer sequences. Advances in neural information processing systems, 33: 0 17283--17297, 2020

  15. [23]

    H2o: Heavy-hitter oracle for efficient generative inference of large language models

    Zhang, Z., Sheng, Y., Zhou, T., Chen, T., Zheng, L., Cai, R., Song, Z., Tian, Y., R \'e , C., Barrett, C., et al. H2o: Heavy-hitter oracle for efficient generative inference of large language models. Advances in Neural Information Processing Systems, 36: 0 34661--34710, 2023

  16. [24]

    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...

Pith tools

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