{ "schema_version": 1, "builder_plan": { "scenario": "A self-hosted assistant is too expensive because its model uses most of the GPU memory.", "change": "Try the paper's supported lower-precision weights on a copy of your serving setup.", "baseline": "The same model at its original precision, on the same GPU.", "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": "Peak memory, task pass rate, latency, and total serving cost at the same context and load.", "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": "2610.29378", "title": "Right On Q: Recovering Reasoning under Quantization via Self-Generated Trajectories", "summary": "OnQ replaces fixed-response teacher supervision in quantization-aware training with on-policy distillation, where the full-precision teacher supervises the quantized student on the student's own generated prefixes, weighted candidate-wise by Dynamic KL. It costs a full training pipeline rather than post-training calibration: student rollout generation plus teacher forward passes per step, which raises compute and implementation complexity beyond block-wise calibration alone. It can fail when the teacher itself is unreliable on student-generated prefixes, when budget limits prevent multiple QAT runs, and its results are demonstrated only on Qwen3 models up to 8B, so transfer to other families or larger scales is unverified.", "rationale": "Recovering reasoning quality in 2-3 bit models matters directly for a 24 GB GPU deployment, but OnQ requires a custom distillation training loop with an available full-precision teacher, so the reader should validate gains on their own model and workload before committing.", "technique": "On-policy distillation with Dynamic KL for QAT", "family": "quantizacao", "recommendation": "testar", "published": "2026-10-21", "as_of": "2026-10-22", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.29378", "independent_implementations": 2, "total_implementations": 3, "stars": 126, "citations": null, "claims": [ { "claim_id": "claim-01", "statement": "OnQ replaces fixed-response teacher supervision in quantization-aware training with on-policy distillation, where the full-precision teacher supervises the quantized student on the student's own generated prefixes, weighted candidate-wise by Dynamic KL. It costs a full training pipeline rather than post-training calibration: student rollout generation plus teacher forward passes per step, which raises compute and implementation complexity beyond block-wise calibration alone. It can fail when the teacher itself is unreliable on student-generated prefixes, when budget limits prevent multiple QAT runs, and its results are demonstrated only on Qwen3 models up to 8B, so transfer to other families or larger scales is unverified.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.29378", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "At 2-bit, OnQ beats every baseline on all nine benchmarks across Qwen3-1.7B/4B/8B, improving the Qwen3-8B nine-benchmark average by 10.26 points over the strongest baseline, with shorter responses and fewer truncations.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.29378", "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": "" }