{ "schema_version": 1, "builder_plan": { "scenario": "A self-hosted model uses more compute or memory than your workload can justify.", "change": "Check for a usable pruned checkpoint and runtime support before attempting model surgery.", "baseline": "The original unpruned model at the same precision, context, and load.", "steps": [ "Locate the paper's implementation and check its license, model support, data needs, and hardware requirements. Stop here if you cannot obtain the required artifacts.", "Save a small set of representative product tasks and expected outcomes. Keep evaluation tasks out of any training or tuning, and choose quality and cost limits before running.", "Run the baseline and the proposed change on identical inputs. Keep other settings fixed and repeat paired runs to expose variation.", "Save per-task outputs, failures, timings, and all usage costs. Review regressions before expanding the test." ], "measure": "Task quality, actual GPU memory, serving throughput, and preparation cost.", "decision": "Keep the change only if it meets your preselected quality and latency limits and improves the metric you care about after setup and operating costs. Otherwise retain the baseline. This is an exploration plan, not a validated recommendation." }, "builder_review": null, "arxiv_id": "2609.18916", "title": "Higher-order pruning of experts in mixture-of-experts language models", "summary": "HOPE replaces first-order expert-importance pruning criteria (e.g., REAP) with a second-order objective that accounts for cooperative interactions between experts when deciding which to remove. It costs additional calibration computation to estimate interaction terms, requires access to a representative calibration set, and does not reduce the non-expert portion of the model, so memory savings are limited to the pruned expert parameters. It can fail if the calibration data does not match the production task distribution, since the cooperative structure it preserves is estimated from that data, and the demonstrated gains were measured on frontier models (up to 122B parameters) that exceed the reader's hardware even after pruning.", "rationale": "The method is directly applicable to shrinking MoE models for memory-constrained deployment and appears superior to existing pruning baselines, but the gain is validated only on very large models and specific calibration sets, so workload-specific testing is required before adoption.", "technique": "Second-order MoE expert pruning (HOPE)", "family": "esparsidade_e_poda", "recommendation": "testar", "published": "2026-09-16", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2609.18916", "independent_implementations": 3, "total_implementations": 3, "stars": 0, "citations": 0, "claims": [ { "claim_id": "claim-01", "statement": "HOPE replaces first-order expert-importance pruning criteria (e.g., REAP) with a second-order objective that accounts for cooperative interactions between experts when deciding which to remove. It costs additional calibration computation to estimate interaction terms, requires access to a representative calibration set, and does not reduce the non-expert portion of the model, so memory savings are limited to the pruned expert parameters. It can fail if the calibration data does not match the production task distribution, since the cooperative structure it preserves is estimated from that data, and the demonstrated gains were measured on frontier models (up to 122B parameters) that exceed the reader's hardware even after pruning.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.18916", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "At 50% expert pruning, HOPE achieves the best average rank (1.58 of 5 methods vs 2.42 for REAP) and gains of up to +6.1% on agentic coding, with the largest advantage at high pruning rates and on agentic workloads.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.18916", "source_page": null, "source_excerpt": "" } ], "exposure_map": [ { "dimension": "quality", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compute", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "latency", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "operations", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compatibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "security", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "data_and_training", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "reproducibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" } ], "risks": [], "minimum_test": [], "open_questions": [], "independent_tests": [], "lab_reproductions": [], "report_available": false, "validation_tier": "unknown", "evidence_tier": "unknown", "infrastructure_basis": "unknown", "equations": [], "equations_status": "not_fetched", "core_kind": "", "equations_fetched_at": "", "equations_source_url": "", "equations_source_sha256": "", "exposures_reviewed_at": "", "exposures_source_sha256": "" }