ThriftAttention recovers 89.1% of the FP16 quality gap versus pure FP4 attention by running only 5% of query-key blocks in FP16 on long-context benchmarks.
TurboAttention: Efficient attention approximation for high throughputs LLMs.arXiv preprint arXiv:2412.08585
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2representative citing papers
Keyless Attention eliminates the key projection, computing softmax(QV^T)V and reducing KV-cache memory by exactly 50%, with comparable or better performance on small-scale models.
citing papers explorer
-
ThriftAttention: Selective Mixed Precision for Long-Context FP4 Attention
ThriftAttention recovers 89.1% of the FP16 quality gap versus pure FP4 attention by running only 5% of query-key blocks in FP16 on long-context benchmarks.
-
Keyless Attention: Value-Space Routing and Value-Only Caching for Efficient Transformers
Keyless Attention eliminates the key projection, computing softmax(QV^T)V and reducing KV-cache memory by exactly 50%, with comparable or better performance on small-scale models.