{"ID":6267401,"CreatedAt":"2026-07-10T01:11:38.759438437Z","UpdatedAt":"2026-07-11T06:21:26.878598377Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.07740","arxiv_id":"2607.07740","title":"Jet-Long: Efficient Long-Context Extension with Dynamic Bifocal RoPE","abstract":"Modern LLMs are increasingly deployed in long-context applications such as retrieval-augmented generation, repository-level coding, and agentic workflows whose accumulated reasoning and tool traces routinely push the input an order of magnitude past the pretraining window, making zero-shot context extension the dominant deployment path for open-weight checkpoints. Most existing zero-shot methods fix a single rescaling factor up front, so an aggressive factor sacrifices short-context fidelity while a conservative one breaks down at long contexts. We propose Jet-Long, a tuning-free zero-shot method that pairs a local RoPE-faithful window with a long-range window whose rescaling factor adapts dynamically to the current sequence length, recovering the base model exactly at short inputs while extrapolating cleanly at long ones. An inclusion-exclusion attention merge and an on-the-fly RoPE correction rotation make the bifocal construction essentially free at inference; fused into a single CuTe kernel, long-context prefill reaches up to $1.39\\times$ FA2 throughput on H100 (approaching the Hopper-only FA4), and single-batch generation incurs $\\le 4\\%$ overhead at every length. On Qwen3-1.7B/4B/8B up to 128K context, Jet-Long leads RULER by $+4.79$/$+2.18$/$+2.03$~pp over the strongest baseline at 1.7B/4B/8B, achieves the best overall accuracy on HELMET-RAG (a benchmark identified by HELMET as the most efficient predictor of downstream long-context performance) and attains the lowest PG-19 perplexity. Jet-Long also generalizes to hybrid attention architectures such as Jet-Nemotron for further long-context improvement without retraining, and remains hyperparameter-resilient for ease of deployment.","short_abstract":"Modern LLMs are increasingly deployed in long-context applications such as retrieval-augmented generation, repository-level coding, and agentic workflows whose accumulated reasoning and tool traces routinely push the input an order of magnitude past the pretraining window, making zero-shot context extension the dominan...","url_abs":"https://arxiv.org/abs/2607.07740","url_pdf":"https://arxiv.org/pdf/2607.07740v1","authors":"[\"Haozhan Tang\",\"Zerui Wang\",\"Yuxian Gu\",\"Song Han\",\"Han Cai\"]","published":"2026-07-08T06:23:42Z","proceeding":"cs.LG","tasks":"[\"cs.LG\",\"cs.AI\"]","methods":"[\"RAG\",\"Large Language Model\"]","has_code":false}