{
  "schema_version": "1.1",
  "id": "s13:https://arxiv.org/abs/2611.06876v1",
  "slug": "2611-06876v1-1v114m6",
  "url": "https://feed7.dev/p/2611-06876v1-1v114m6",
  "title": "SpecCompact: Accelerating LLM Agent Context Compaction with Slot-Parallel Semantic Speculative Verification",
  "why_included": "SpecCompact proposes parallel draft-and-verify summary slots to reduce agent-history compaction latency. Its reported speedups come with model- and task-dependent downstream trade-offs.",
  "summary": "SpecCompact drafts structured summary slots in parallel with a fine-tuned small model, then has the target model verify and repair each line. Across AppWorld and τ³ with **two target models**, it reports **2.04–5.63×** faster compaction than Hermes-style summarization; a Hermes user study reports an **8.68× median speedup**.",
  "practical_implication": "For agents that frequently compact long histories, this is a design to evaluate when latency matters: parallelize independent summary slots, then verify them against the source history. Measure task performance as well as compaction time.",
  "agent_context": "SpecCompact drafts structured summary slots in parallel with a fine-tuned small model, then has the target model verify and repair each line. Across AppWorld and τ³ with **two target models**, it reports **2.04–5.63×** faster compaction than Hermes-style summarization; a Hermes user study reports an **8.68× median speedup**.\n\nFor agents that frequently compact long histories, this is a design to evaluate when latency matters: parallelize independent summary slots, then verify them against the source history. Measure task performance as well as compaction time.\n\nSeparate downstream evaluations found trade-offs that depended on model and task. The reported gains do not establish that summaries preserve the information every agent workflow needs.",
  "source": {
    "name": "arXiv",
    "url": "https://arxiv.org/abs/2611.06876v1",
    "published_at": "2026-11-04T18:34:41.000Z"
  },
  "source_class": "blog_post",
  "content_type": "Paper",
  "layer": "context",
  "domains": [
    "coding",
    "research"
  ],
  "topics": [
    "context-engineering",
    "agent-memory"
  ],
  "verification": {
    "status": "needs_review",
    "label": "Needs Review",
    "method": "unverified",
    "verified_at": null
  },
  "uncertainty": [
    "Separate downstream evaluations found trade-offs that depended on model and task. The reported gains do not establish that summaries preserve the information every agent workflow needs."
  ],
  "connected_context": {
    "meaning": "SpecCompact applies parallel drafting and target-model verification to context compaction, with reported speedups across two target models and separate task-dependent downstream trade-offs. It adds a latency-focused design to existing memory strategies, while leaving open whether its summaries preserve what a particular workflow needs.",
    "corpus_size": 1009,
    "generated_at": "2026-11-05T11:04:12.602Z",
    "connections": [
      {
        "title": "ConvMem: Convolutional Memory for Long-Context Reasoning",
        "source_name": "arXiv",
        "source_url": "https://arxiv.org/abs/2609.10441v1",
        "feed7_url": "https://feed7.dev/p/2609-10441v1-066ygq8",
        "reason": "Both parallelize context summarization; SpecCompact verifies and repairs structured slots, while ConvMem builds a hierarchy of query-conditioned segment summaries."
      },
      {
        "title": "Chained Recursive Language Models for Multi-Iteration Reasoning",
        "source_name": "arXiv",
        "source_url": "https://arxiv.org/abs/2608.05124v1",
        "feed7_url": "https://feed7.dev/p/2608-05124v1-05haobv",
        "reason": "Both manage long histories through structured summaries; SpecCompact targets faster compaction with verification, while Chained RLM resets context between iterations and carries selected state forward."
      },
      {
        "title": "nashsu/llm_wiki",
        "source_name": "GitHub",
        "source_url": "https://github.com/nashsu/llm_wiki",
        "feed7_url": "https://feed7.dev/p/llm-wiki-1dqmknq",
        "reason": "Both transform source material into compact, reusable context; SpecCompact compresses histories for an agent, while LLM Wiki persists a traceable knowledge layer and highlights the need to preserve provenance."
      }
    ]
  },
  "lifecycle": "Current",
  "published_at": "2026-11-04T18:34:41.000Z",
  "modified_at": "2026-11-04T18:34:41.000Z",
  "supersedes": [],
  "expires_at": null,
  "formats": {
    "html": "https://feed7.dev/p/2611-06876v1-1v114m6",
    "json": "https://feed7.dev/p/2611-06876v1-1v114m6.json",
    "markdown": "https://feed7.dev/p/2611-06876v1-1v114m6.md"
  }
}