Transformers with Selective Access to Early Representations

• URL: http://arxiv.org/abs/2605.03953v2
• Date: Wed, 06 May 2026 01:32:44 GMT
• Title: Transformers with Selective Access to Early Representations
• Abstract summary: We introduce Selective Access Transformer (SATFormer), which preserves the first-layer value pathway while controlling access with a context-dependent gate.<n>SATFormer consistently improves validation loss and zero-shot accuracy over the static value-residual and Transformer baselines.<n> Gate analyses suggest sparse, depth-dependent, head-specific, and category-sensitive access patterns.
• Score: 6.554898376530473
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Several recent Transformer architectures expose later layers to representations computed in the earliest layers, motivated by the observation that low-level features can become harder to recover as the residual stream is repeatedly transformed through depth. The cheapest among these methods add static value residuals: learned mixing coefficients that expose the first-layer value projection V_1 uniformly across tokens and heads. More expressive dense or dynamic alternatives recover finer-grained access, but at higher memory cost and lower throughput. The usefulness of V_1 is unlikely to be constant across tokens, heads, and contexts; different positions plausibly require different amounts of access to early lexical or semantic information. We therefore treat early-representation reuse as a retrieval problem rather than a connectivity problem, and introduce Selective Access Transformer (SATFormer), which preserves the first-layer value pathway while controlling access with a context-dependent gate. Across models from 130M to 1.3B parameters, SATFormer consistently improves validation loss and zero-shot accuracy over the static value-residual and Transformer baselines. Its strongest gains appear on retrieval-intensive benchmarks, where it improves over static value residuals by approximately 1.5 average points, while maintaining throughput and memory usage close to the baseline Transformer. Gate analyses suggest sparse, depth-dependent, head-specific, and category-sensitive access patterns, supporting the interpretation that SATFormer learns selective reuse of early representations rather than uniform residual copying. Our code is available at https://github.com/SkyeGunasekaran/SATFormer.

Related papers

• T^2MLR: Transformer with Temporal Middle-Layer Recurrence [28.818755729513793]
  We develop a transformers-based latent reasoning architecture that fuses a cached middle layer representation from the previous token directly into an earlier layer of the current token position.<n>Applying recurrence to only a localized middle-layer block (as little as 20% of the network) often outperforms full-layer recurrence.<n>These results suggest that effective latent reasoning in Transformers does not require looping over all layers as in previous works, but can instead emerge more strongly from targeted middle-layer recurrence.
  arXiv  Detail & Related papers  (2026-07-16T16:33:59Z)
• Latent Recurrent Transformer: Architecture Exploration, Training Strategies, and Scaling Behavior [107.2098567818173]
  Latent Recurrent Transformer (LRT) is a lightweight augmentation of autoregressive transformers.<n>LRT reuses a high-level source-layer hidden state from the previous token as recurrent memory for the next token.
  arXiv  Detail & Related papers  (2026-05-26T10:10:26Z)
• The Recurrent Transformer: Greater Effective Depth and Efficient Decoding [48.9323408950142]
  We introduce the Recurrent Transformer, a simple architectural change where each layer attends to key-value pairs computed off its own activations.<n>We show that the architecture can emulate both (i) a conventional Transformer and (ii) token-to-gressive recurrent updates under mild assumptions.
  arXiv  Detail & Related papers  (2026-04-23T02:12:58Z)
• T1: One-to-One Channel-Head Binding for Multivariate Time-Series Imputation [6.002503434201551]
  Imputing missing values in time series is challenging under diverse missing patterns and heavy missingness.<n>We introduce T1 (Time series imputation with 1-to-1 channel-head binding), a CNN-Transformer hybrid architecture that achieves robust imputation.<n>T1 achieves state-of-the-art performance, reducing MSE by 46% on average compared to the second-best baseline.
  arXiv  Detail & Related papers  (2026-02-24T16:03:46Z)
• MeSH: Memory-as-State-Highways for Recursive Transformers [23.995570647573484]
  Recursive models with fewer parameters often lag behind non-recursive counterparts under matched compute.<n>By probing hidden states, we trace this performance gap to two primary bottlenecks.<n>We introduce a Memory-as-State-Highways scheme, which externalizes state management into an explicit memory buffer.
  arXiv  Detail & Related papers  (2025-10-09T03:23:38Z)
• You Do Not Fully Utilize Transformer's Representation Capacity [4.753535328327317]
  Layer-Integrated Memory (LIMe) is a lightweight extension that learns per-head, per-layer routing weights to integrate representations from all previous layers with negligible overhead.<n>LIMe consistently achieves faster convergence, lower perplexity per FLOP, and substantial accuracy improvements on synthetic tasks.
  arXiv  Detail & Related papers  (2025-02-13T12:00:50Z)
• Sparser Training for On-Device Recommendation Systems [50.74019319100728]
  We propose SparseRec, a lightweight embedding method based on Dynamic Sparse Training (DST) It avoids dense gradients during backpropagation by sampling a subset of important vectors.
  arXiv  Detail & Related papers  (2024-11-19T03:48:48Z)
• Value Residual Learning [13.88704205151734]
  This paper introduces ResFormer, a novel architecture that enhances information flow by incorporating value residual connections.<n>It achieves equivalent validation loss with 16.11% fewer model parameters and 20.3% less training data compared to Transformer.
  arXiv  Detail & Related papers  (2024-10-23T14:15:07Z)
• A Fast Post-Training Pruning Framework for Transformers [74.59556951906468]
  Pruning is an effective way to reduce the huge inference cost of large Transformer models. Prior work on model pruning requires retraining the model. We propose a fast post-training pruning framework for Transformers that does not require any retraining.
  arXiv  Detail & Related papers  (2022-03-29T07:41:11Z)
• Improving language models by retrieving from trillions of tokens [50.42630445476544]
  We enhance auto-regressive language models by conditioning on document chunks retrieved from a large corpus. With a $2$ trillion token database, our Retrieval-Enhanced Transformer (RETRO) obtains comparable performance to GPT-3 and Jurassic-1 on the Pile.
  arXiv  Detail & Related papers  (2021-12-08T17:32:34Z)
• Memory-efficient Transformers via Top-$k$ Attention [23.672065688109395]
  In this work, we propose a simple yet highly accurate approximation for vanilla attention. We process the queries in chunks, and for each query, compute the top-$k$ scores with respect to the keys. We show our approach leads to accuracy that is nearly-identical to vanilla attention in multiple setups including training from scratch, fine-tuning, and zero-shot inference.
  arXiv  Detail & Related papers  (2021-06-13T02:30:23Z)
• GradInit: Learning to Initialize Neural Networks for Stable and Efficient Training [59.160154997555956]
  We present GradInit, an automated and architecture method for initializing neural networks. It is based on a simple agnostic; the variance of each network layer is adjusted so that a single step of SGD or Adam results in the smallest possible loss value. It also enables training the original Post-LN Transformer for machine translation without learning rate warmup.
  arXiv  Detail & Related papers  (2021-02-16T11:45:35Z)
• Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing [112.2208052057002]
  We propose Funnel-Transformer which gradually compresses the sequence of hidden states to a shorter one. With comparable or fewer FLOPs, Funnel-Transformer outperforms the standard Transformer on a wide variety of sequence-level prediction tasks.
  arXiv  Detail & Related papers  (2020-06-05T05:16:23Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.