The paper introduces First Passage Attention (FPA), a new attention mechanism that replaces the traditional summation of decay contributions with a path‑extrema approach, lowering a key’s logit based on the deepest dip of a learned scalar field over the intervening span. FPA generalises the Forgetting Transformer (FoX) and, when the field and levels share a learned walk, naturally forms a stack of alive keys. Experiments show that a constant‑dimension FPA head can realise any subset of keys left alive by subsequent dips, outperforming softmax heads and FoX baselines in synthetic tasks, and matching FoX perplexity while achieving perfect passkey retrieval at eight times the training length in a 110M‑parameter language model.
arXiv:2607. 12204v1 Announce Type: new Abstract: Attention can be viewed as an online learner over context, yet existing test-time memories cannot certify that dropping a token leaves outputs unchanged or delete its influence outright.
arXiv:2610.09835v1 Announce Type: new Abstract: Continual pre-training and fine-tuning in Large Language Models (LLMs) inevitably induce catastrophic forgetting, typically mitigated by replay using o...
The paper introduces ADU, a fine‑grained training framework that unlearns sensitive information from large language models by decoupling contextual attention pathways instead of erasing tokens. ADU exploits the distinction between local and global attention heads to identify and suppress attention paths that retrieve persistent sensitive anchors, while preserving local‑attention structure and overall language modeling performance. Evaluation on the TOFU and WMDP benchmarks shows ADU achieves superior forget quality (0.93 on TOFU) and retains 92.9% of model utility compared to 81.9% for existing baselines, with fewer side effects in benign contexts.
arXiv:2605. 25475v2 Announce Type: replace-cross Abstract: Large Language Models (LLMs) are increasingly expected to operate over long contexts, yet standard softmax attention incurs a KV cache that grows linearly with sequence length, quickly becoming the bottleneck for long context inference.
arXiv:2606. 01294v1 Announce Type: cross Abstract: Linear attention reduces the quadratic cost of softmax attention by maintaining a recurrent fast-weight state, but it consistently lags on in-context retrieval and long-context tasks.
arXiv:2505. 23666v3 Announce Type: replace-cross Abstract: The per-token cost of transformer inference scales with context length, preventing its application to lifelong in-context learning.
arXiv:2601. 18699v2 Announce Type: replace Abstract: Sequential fine-tuning of Large Language Models (LLMs) adaptation to target tasks often triggers catastrophic forgetting, where the acquisition of novel target skills degrades ancestral capabilities.
The paper investigates how gating the value pathway in attention mechanisms provides two missing capabilities of softmax attention: abstention and noise filtering. Experiments on models ranging from 10M to 350M parameters show that abstention benefits smaller models while noise filtering becomes more advantageous as models scale, and that combining both primitives yields the best performance across all sizes. The authors also demonstrate that the gates effectively suppress interference and that each gate type has a distinct blind spot, all while adding negligible parameters and preserving compatibility with key‑value caching.
arXiv:2606. 25156v1 Announce Type: new Abstract: Modern large language models based on softmax scaled-dot-product attention are constrained by their training sequence length: as the key-value sequence grows, softmax probability mass can dilute across a wider distribution, inducing activation shift and long-context performance collapse.
The paper investigates how attention dynamics evolve across recurrent depth in language models, finding that attention support stabilizes early while hidden states and outputs take longer. It proposes WISE, a training‑free method that uses full attention in early steps and then reuses the discovered sparse working set for later steps, preserving performance on multi‑hop QA tasks. Experiments show that WISE maintains quality up to 2K context, offers measurable speedups, and highlights the importance of recurrent discovery of attention support.
arXiv:2609.25802v1 Announce Type: new Abstract: We introduce latest exact match attention (LEMA), an attention variant for transformers where queries and keys are binarized and each query attends onl...
The paper introduces Universal Attention, a transformer architecture that unifies and extends decay-based KV‑cache pruning mechanisms while preserving RoPE embeddings and Softmax attention. It achieves highly expressive, end‑to‑end trainable pruning that adaptively removes low‑contribution tokens, yielding up to 10× compression on standard datasets and 25× compression at 16k context length without sacrificing downstream performance. The approach outperforms existing state‑of‑the‑art baselines and unpruned models in both compression and generalization.