The following papers were recommended by the Semantic Scholar API
\n- \n
- HAMburger: Accelerating LLM Inference via Token Smashing (2025) \n
- Beyond In-Context Learning: Aligning Long-form Generation of Large Language Models via Task-Inherent Attribute Guidelines (2025) \n
- MMLongBench: Benchmarking Long-Context Vision-Language Models Effectively and Thoroughly (2025) \n
- Scaling Context, Not Parameters: Training a Compact 7B Language Model for Efficient Long-Context Processing (2025) \n
- MateICL: Mitigating Attention Dispersion in Large-Scale In-Context Learning (2025) \n
- Compress, Gather, and Recompute: REFORMing Long-Context Processing in Transformers (2025) \n
- Efficient Pretraining Length Scaling (2025) \n
Please give a thumbs up to this comment if you found it helpful!
\nIf you want recommendations for any Paper on Hugging Face checkout this Space
\n You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: \n\n@librarian-bot\n\t recommend
Cartridges: Lightweight and general-purpose long context representations via self-study
Abstract
Training a smaller, offline KV cache (Cartridge) with a context-distillation objective (self-study) for large language models reduces serving costs, matches ICL performance, and extends effective context length.
Large language models are often used to answer queries grounded in large text corpora (e.g. codebases, legal documents, or chat histories) by placing the entire corpus in the context window and leveraging in-context learning (ICL). Although current models support contexts of 100K-1M tokens, this setup is costly to serve because the memory consumption of the KV cache scales with input length. We explore an alternative: training a smaller KV cache offline on each corpus. At inference time, we load this trained KV cache, which we call a Cartridge, and decode a response. Critically, the cost of training a Cartridge can be amortized across all the queries referencing the same corpus. However, we find that the naive approach of training the Cartridge with next-token prediction on the corpus is not competitive with ICL. Instead, we propose self-study, a training recipe in which we generate synthetic conversations about the corpus and train the Cartridge with a context-distillation objective. We find that Cartridges trained with self-study replicate the functionality of ICL, while being significantly cheaper to serve. On challenging long-context benchmarks, Cartridges trained with self-study match ICL performance while using 38.6x less memory and enabling 26.4x higher throughput. Self-study also extends the model's effective context length (e.g. from 128k to 484k tokens on MTOB) and surprisingly, leads to Cartridges that can be composed at inference time without retraining.
Community
When we put lots of text (e.g. a whole code repo) into a language model’s context, generation cost soars because of the KV cache’s size. What if we trained a smaller KV cache for our documents offline? Using a test-time training recipe called self-study, we show that this simple idea can improve throughput by 26x while maintaining quality.
This is an automated message from the Librarian Bot. I found the following papers similar to this paper.
The following papers were recommended by the Semantic Scholar API
- HAMburger: Accelerating LLM Inference via Token Smashing (2025)
- Beyond In-Context Learning: Aligning Long-form Generation of Large Language Models via Task-Inherent Attribute Guidelines (2025)
- MMLongBench: Benchmarking Long-Context Vision-Language Models Effectively and Thoroughly (2025)
- Scaling Context, Not Parameters: Training a Compact 7B Language Model for Efficient Long-Context Processing (2025)
- MateICL: Mitigating Attention Dispersion in Large-Scale In-Context Learning (2025)
- Compress, Gather, and Recompute: REFORMing Long-Context Processing in Transformers (2025)
- Efficient Pretraining Length Scaling (2025)
Please give a thumbs up to this comment if you found it helpful!
If you want recommendations for any Paper on Hugging Face checkout this Space
You can directly ask Librarian Bot for paper recommendations by tagging it in a comment:
@librarian-bot
recommend
Models citing this paper 0
No model linking this paper
Datasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 0
No Space linking this paper