mlsanityData to 14 Oct 2026

Governance Records as Supervision: Verifier-Selected Self-Training for Structured Workflow Repair

arXiv 2608.18324 ↗First posted 18 Aug 2026cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

Machine-verifiable workflows produce governance records linking a task contract, model attempt, verifier decision, accepted output, and target origin. We test whether verifier-admitted outputs can supervise a bounded model by consolidating occasional or expensive capability into reliable one-shot execution. On fresh, structure-disjoint PlanBench replanning cases, Qwen3-14B thinking produced 24 plans admitted by independently authored VAL. They trained the same checkpoint for non-thinking execution, without oracle targets or a stronger teacher. VAL acceptance rose from 1/80 to 57/80. A prospective replication held targets, model revision, recipe, and evaluation corpus fixed across eight LoRA seeds and three inference realizations per seed. Every seed produced a clear lift: adapters reached 45/80 to 70/80 against 1/80 for every matched base report; the exact seed-level sign-flip test gave p=0.0078125. Target-selection performance was less stable. An initial matched seed gave 102/160 accepted plans after VAL selection versus 69/160 after blinded model self-selection. Across eight prospective seeds, the contrast was seed-dependent, included one clear reverse seed, and did not replicate (p=0.3672). VAL also had a positive descriptive aggregate over schema-only selection but failed its preregistered seed-level reliability gate (p=0.0703). The verifier remains the admission authority; no reliable downstream capability advantage of semantic selection is established. A complementary Phi arm supports stronger-teacher distillation. Earlier synthetic studies bound teachability, cumulative learning, transfer, and stopping. The evidence supports robust consolidation of one fixed, machine-checkable capability, not arbitrary planning, enterprise validity, or unrestricted self-improvement.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 15 of its 27 references that are in the map. First posted on arXiv 18 Aug 2026.

Builds on

15 of 27 references.

References · 15

  1. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
  2. 2015Distilling the Knowledge in a Neural Networkcited by 10,324
  3. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  4. 2023Reflexion: Language Agents with Verbal Reinforcement Learningcited by 4,276
  5. 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 2,724
  6. 2023Large Language Models Cannot Self-Correct Reasoning Yetcited by 1,017
  7. 2023Teaching Large Language Models to Self-Debugcited by 988
  8. 2024Self-Play Fine-Tuning Converts Weak Language Models to Strong Language Modelscited by 557
  9. 2023Scaling Relationship on Learning Mathematical Reasoning with Large Language Modelscited by 427
  10. 2022PlanBench: An Extensible Benchmark for Evaluating Large Language Models on Planning and Reasoning about Changecited by 382
  11. 2022CodeRL: Mastering Code Generation through Pretrained Models and Deep Reinforcement Learningcited by 358
  12. 2023On the Planning Abilities of Large Language Models : A Critical Investigationcited by 312
3 more
  1. 2023LEVER: Learning to Verify Language-to-Code Generation with Executioncited by 239
  2. 2024On the Self-Verification Limitations of Large Language Models on Reasoning and Planning Taskscited by 142
  3. 2026Correct Is Not Governed: Provenance Integrity in Agentic Workflowscited by 2

Built on by

Not cited yet.