{ "schema_version": 1, "builder_plan": { "scenario": "You already need task-specific training and want to reduce its memory or compute cost.", "change": "Change only the training technique on a small controlled run after checking prerequisites.", "baseline": "Your current training recipe with the same model, data split, and evaluation.", "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": "Held-out quality, peak memory, training time, and full compute cost to the target quality.", "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": "2610.00949", "title": "PG-SFT: Balancing Capability Acquisition and Retention in Offline Agent Fine-Tuning", "summary": "PG-SFT replaces uniform token-by-token supervised fine-tuning on offline agent trajectories with a token-wise objective that scales supervision strength by turn-level information gain, rather than relying on KL penalties or update-magnitude constraints. It costs a slight reduction in target-task performance and adds the complexity of computing per-turn information gain over the training trajectories. It can fail if the information-gain signal is noisy or misestimated on the reader's trajectory data, or if the retention-acquisition balance it was tuned for does not transfer to different base models or tool-use domains.", "rationale": "The method directly addresses a real production problem (capability regression after agent SFT) and is feasible on a single 24 GB GPU, but its benefit depends on the reader's specific trajectory data and base model, so it requires workload-specific validation before adoption.", "technique": "Privilege-Guided SFT with turn-level information gain weighting", "family": "treino_eficiente", "recommendation": "testar", "published": "2026-10-01", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.00949", "independent_implementations": 2, "total_implementations": 3, "stars": 135, "citations": 0, "claims": [ { "claim_id": "claim-01", "statement": "PG-SFT replaces uniform token-by-token supervised fine-tuning on offline agent trajectories with a token-wise objective that scales supervision strength by turn-level information gain, rather than relying on KL penalties or update-magnitude constraints. It costs a slight reduction in target-task performance and adds the complexity of computing per-turn information gain over the training trajectories. It can fail if the information-gain signal is noisy or misestimated on the reader's trajectory data, or if the retention-acquisition balance it was tuned for does not transfer to different base models or tool-use domains.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.00949", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "PG-SFT substantially reduces distributional drift and non-target capability degradation relative to standard SFT, at the cost of a slight drop in target-benchmark performance; no multiplicative factor is reported in the abstract.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.00949", "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": "" }