arXiv cs.CL· Roberto I. Ono Filho·· 13 小时前AI 评分22
推理时干预与权重整合在开放式生成中能带来什么：让均值靠向已知最优而非超越
Moving the Mean Toward the Known Good, Not Beyond It: What Inference-Time Interventions and Weight Consolidation Buy in Open-Ended Generation
AI 导读
在在线装箱任务上对通过验证的候选做 LoRA 整合训练，可使模型在留出变体上的输出向经典启发式水平靠拢（超额 -1.7 点，p=0.008；相对随机整合对照 -3.1，p=0.004），但最佳观测候选仅收敛到经典启发式水平、未再超越。
正文 · 原文
Abstract:What does a generation loop gain from learning on its own verified successes? In cycles of generate, verify, select and LoRA-consolidate on online bin packing, training on value-filtered candidates shifts what the model writes on held-out variants toward value (-1.7 points of excess, p=0.008; -3.1 against a random-consolidation control, p=0.004) while the best observed candidate converges to the classic heuristic's level and no further. A confirmation battery replicates the whole procedure three times, with fresh seeds and a never-consulted held-out set read exactly once: the mean was nearly identical in all three lineages (-2.0, -1.8, -1.9), and after aggregating within held-out variant all seven evaluable variants favored consolidation (p=0.008). The best observed candidate moved to the classic heuristic's level, exactly (0.021028, the lowest variant's classic level, in all three lineages and both arms); attract never went beyond its variant's classic level, while the random control did in 7 of 24 lineage-variant cells. A matched SFT-only control shows the supervised anchor, not repulsion from bad candidates, does the concentrating (96% of candidates land exactly at the classic heuristic's level). On far families the tails cut both ways: consolidation lowers the per-candidate rate of better-than-classic candidates (10% to 3.9%) while its larger production yields more such candidates absolutely (5 against 1, on few events). As motivation we report the inference-time ledger that led here: a model-written schematic recap buys judged document integration and nothing buys development; a verifier written into the stream is imitated, 16.4 fabricated verdict lines per notebook. In this arrangement (an 8-bit model in MLX, five cycles), mean quality among valid candidates can be bought and replicated; under attraction the observed best goes to the classic and, so far, never beyond it.
| Comments: | Version 2: corrections after an audit against the records, including the replication's beyond-classic count for the random control (listed in the paper). 14 pages. Code: this https URL. Data: Zenodo https://doi.org/10.5281/zenodo.22714138 and https://doi.org/10.5281/zenodo.23163299%3B this version also at https://doi.org/10.5281/zenodo.23163706 |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| ACM classes: | I.2.7; I.2.6 |
| Cite as: | arXiv:2608.28886 [cs.CL] |
| (or arXiv:2608.28886v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2608.28886 arXiv-issued DOI via DataCite |
Submission history
From: Roberto I Ono Filho [view email]
[v1]
Fri, 28 Aug 2026 21:41:17 UTC (211 KB)
[v2]
Thu, 22 Oct 2026 12:05:11 UTC (213 KB)
来源：arXiv cs.CL · arxiv.org