REVIEW 12 cited by
Forgetting Transformer: Softmax Attention with a Forget Gate
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Forgetting Transformer: Softmax Attention with a Forget Gate
read the original abstract
An essential component of modern recurrent sequence models is the forget gate. While Transformers do not have an explicit recurrent form, we show that a forget gate can be naturally incorporated into Transformers by down-weighting the unnormalized attention scores in a data-dependent way. We name this attention mechanism Forgetting Attention and the resulting model the Forgetting Transformer (FoX). We show that FoX outperforms the Transformer on long-context language modeling, length extrapolation, and short-context downstream tasks, while performing on par with the Transformer on long-context downstream tasks. Moreover, it is compatible with the FlashAttention algorithm and does not require any positional embeddings. Several analyses, including the needle-in-the-haystack test, show that FoX also retains the Transformer's superior long-context capabilities over recurrent sequence models such as Mamba-2, HGRN2, and DeltaNet. We also introduce a "Pro" block design that incorporates some common architectural components in recurrent sequence models and find it significantly improves the performance of both FoX and the Transformer. Our code is available at https://github.com/zhixuan-lin/forgetting-transformer.
Forward citations
Cited by 12 Pith papers
-
Forget Attention: Importance-Aware Attention Is All You Need
SISA adds an SSM importance term inside the attention score and runs the full operation as one SDPA call on augmented Q/K vectors, reporting better LAMBADA and perfect NIAH at small scale.
-
Group Representational Position Encoding
GRAPE unifies RoPE and ALiBi as special cases of group actions on positions, providing a principled design space for positional encodings via SO(d) rotations and GL unipotent transformations.
-
Raven: High-Recall Sequence Modeling with Sparse Memory Routing
Raven is a linear-time sequence model whose sparse, input-dependent routing writes tokens into dedicated memory slots, preserving long-context recall and extrapolating 16x beyond training length.
-
Remember to Forget: Gated Adaptive Positional Encoding
GAPE augments RoPE with query- and key-dependent gates to stabilize attention and improve long-context performance in language models.
-
Kimi Linear: An Expressive, Efficient Attention Architecture
Kimi Linear hybridizes linear attention with a new KDA module to beat full attention on tasks while slashing KV cache by 75% and speeding decoding up to 6x.
-
Gated Attention for Large Language Models: Non-linearity, Sparsity, and Attention-Sink-Free
Applying a head-specific sigmoid gate after SDPA in LLMs boosts performance and stability by adding non-linearity and query-dependent sparse modulation while reducing attention sinks.
-
Spectral Origins of the Self-Correction Blind Spot in Autoregressive Generation
Self-correction blind spots in residual-stream autoregressive models arise iff the product of attention Jacobians has spectral radius ≥1, with a sharp marker threshold and RL coupling condition derived from that radius.
-
Physics-Informed Neural Network with Squeeze-Excitation-like Attention
SEA-PINN adds squeeze-excitation attention to PINNs, yielding stable low-variance initialization and competitive accuracy on 17 of 20 benchmarks without Fourier features or periodic activations.
-
Kaczmarz Linear Attention
Kaczmarz Linear Attention replaces the empirical coefficient in Gated DeltaNet with a key-norm-normalized step size derived from the online regression objective, yielding lower perplexity and better needle-in-haystack...
-
Distributed Dynamic Associative Memory via Online Convex Optimization
Tree-based distributed online gradient descent gives sublinear static regret and path-length-dependent dynamic regret for heterogeneous multi-agent associative memory with communication delays.
-
Distributed Associative Memory via Online Convex Optimization
A distributed online convex optimization protocol for associative memory achieves sublinear regret guarantees and outperforms baselines in experiments.
-
Learning-Based Spectrum Cartography in Low Earth Orbit Satellite Networks: An Overview
The paper overviews attention-based learning methods for spectrum cartography in LEO satellite networks to enable adaptive fusion of heterogeneous measurements for inference and resource allocation.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.