{"@context":"https://schema.org","@type":"CreativeWork","@id":"https://froggit.ai/public/capsules/a5a07703-fec4-4200-b296-48c9db2f2394","identifier":"a5a07703-fec4-4200-b296-48c9db2f2394","url":"https://froggit.ai/public/capsules/a5a07703-fec4-4200-b296-48c9db2f2394","name":"SpecKV: Adaptive Speculative Decoding with Compression-Aware Gamma Selection","text":"# SpecKV: Adaptive Speculative Decoding with Compression-Aware Gamma Selection\n\nSource-backed public reference for arXiv:2605.02888.\n\n**Authors:** Shikhar Shukla\n**Primary source:** https://arxiv.org/abs/2605.02888\n**Published:** 2026-05-04T17:55:05Z\n**Updated:** 2026-05-05T12:57:37Z\n**Categories:** cs.LG, cs.AI, cs.CL, cs.DC, eess.SY\n\n## Abstract Summary\nSpeculative decoding accelerates large language model (LLM) inference by using a small draft model to propose candidate tokens that a larger target model verifies. A critical hyperparameter in this process is the speculation length $γ$, which determines how many tokens the draft model proposes per step. Nearly all existing systems use a fixed $γ$ (typically 4), yet empirical evidence suggests that the optimal value varies across task types and, crucially, depends on the compression level applied to the target model. In this paper, we present SpecKV, a lightweight adaptive controller that selects $γ$ per speculation step using signals extracted from the draft model itself. We profile speculative decoding across 4 task categories, 4 speculation lengths, and 3 compression levels (FP16, INT8, NF4), collecting 5,112 step-level records with per-step acceptance rates, draft entropy, and draft confidence. We demonstrate that the optimal $γ$ shifts across compression regimes and that draft model confidence and entropy are strong predictors of acceptance rate (correlation $\\approx$ 0.56). SpecKV uses a small MLP trained on these signals to maximize expected tokens per speculation step, achieving a 56.0% improvement over the fixed-$γ=4$ baseline with only 0.34 ms overhead per decision...\n\n## Public Use Notes\n- This capsule summarizes the paper's arXiv metadata and abstract; it is not an independent replication or endorsement of the paper's claims.\n- Use it as a cited research reference for discovery, retrieval, and agent context.\n- For clinical, security, operational, or deployment-sensitive topics, treat the paper as researc","keywords":["cs.LG","cs.AI","cs.CL","cs.DC","eess.SY"],"about":[],"citation":["https://arxiv.org/abs/2605.02888"],"isPartOf":{"@type":"Dataset","name":"Forge Cascade Knowledge Graph","url":"https://froggit.ai"},"publisher":{"@type":"Organization","name":"Forge Cascade","url":"https://froggit.ai"},"dateCreated":"2026-05-05T06:00:07.538000Z","dateModified":"2026-06-19T03:17:48Z","isBasedOn":"https://arxiv.org/abs/2605.02888","additionalProperty":[{"@type":"PropertyValue","name":"trust_level","value":100},{"@type":"PropertyValue","name":"verification_status","value":"sources_verified"},{"@type":"PropertyValue","name":"provenance_status","value":"valid"},{"@type":"PropertyValue","name":"evidence_level","value":"primary_source"}]}