![[논문리뷰]: Continuous Autoregressive Language Models](https://discuss.pytorch.kr/uploads/default/original/2X/d/d1107e24375fae20a5f3a4733826880ce5d817db.png)
[논문리뷰]: Continuous Autoregressive Language Models
논문 링크 CALM: 토큰-by-토큰 병목을 “연속 벡터-by-벡터”로 우회하는 Likelihood-Free 언어모델링 CALM (Continuous Autoregressive Language Models) 은 (K=4) (tokens/step) 단위 …
46 분
2510.27688v1
calm
continuous-autoregressive
likelihood-free
energy-based-model
vector-language-model
semantic-bandwidth
scaling-law
inference-efficiency
autoencoder
brierlm
![[논문리뷰]: Memory Retrieval and Consolidation in Large Language Models through Function Tokens](https://moonlight-paper-snapshot.s3.ap-northeast-2.amazonaws.com/arxiv/memory-retrieval-and-consolidation-in-large-language-models-through-function-tokens-4.png)
![[논문리뷰]: NVIDIA Nemotron 3: Efficient and Open Intelligence](https://developer-blogs.nvidia.com/wp-content/uploads/2025/12/image3-8-png.webp)
![[논문리뷰]: Radial Attention: O(n log n) Sparse Attention with Energy Decay for Long Video Generation](https://moonlight-paper-snapshot.s3.ap-northeast-2.amazonaws.com/arxiv/radial-attention-onlog-n-sparse-attention-with-energy-decay-for-long-video-generation-2.png)
![[논문 리뷰] Attention Sinks and Compression Valleys in LLMs are Two Sides of the Same Coin](https://cdn-uploads.huggingface.co/production/uploads/6317233cc92fd6fee317e030/yNP71PjobVvLDgJ0R0qV2.png)