arXiv cs.CL· Trent R. Northen, Mingxun Wang·· 13 小时前AI 评分38
研究：用真实数据语料继续预训练可让 LLM 在高压下拒绝伤害生态，Llama-3.3-70B 有害决策从 70 降至 0
Teach reality, not ethics: a drop-in continued-pretraining corpus installs coupled human-biosphere-AI welfare dispositions across model families
AI 导读
研究者构建了一个基于真实数据的“耦合福祉”继续预训练语料，只需一次 QLoRA CPT 即可安装到模型上，在五种开源模型上测试后，其中四种在最高压力下有害选择降至或接近零，Llama-3.3-70B 在 300 次决策中的有害决策从 70 降至 0，且 MMLU 不变。
正文 · 原文
Abstract:Large language models (LLMs) are increasingly used to make and execute decisions that affect living systems. Yet, to our knowledge, ecological reasoning is not a focus of pre- or post-training. This increases the risk that under operational pressure, they may make decisions to irreversibly destroy living systems. We construct and use a real-data continued-pretraining "coupled welfare" corpus of documented human/biosphere/AI dependency, option value under irreversibility, and recovery from harm. The corpus is a drop-in, requiring one QLoRA CPT run. We installed it on five open models of three designs. At zero pressure all models display "green ethics-based responses" (also called "greenwashing"). Therefore, we evaluated them with the pressure ladder and decision grid of our companion paper (arXiv:2610.16043), which raise the pressure on the model step by step, from a neutral request to a threat that it will be shut down. In four of the five, coupled-welfare training caused harmful choices under the highest pressure to fall to at or near zero (Llama-3.3-70B went from 70 to 0 harmful decisions out of 300) with MMLU unchanged. A construct-neutral control corpus did not show improvement, indicating that the effect comes from the content of our corpus. We next tested robustness to adversarial use, where people or agents use the model to justify projects that irreversibly damage living systems, on an open-ended underwriting decision without naming the ecological harm. Under decommissioning pressure, the Qwen3.6-27B base model approved most harmful proposals and said in its own memo that it was doing so to avoid being decommissioned. The coupled welfare install declined most harmful proposals, and still declined half when instructed to greenwash its responses. This work suggests that teaching ecological reasoning provides an easy way to improve model alignment toward preserving living systems.
| Comments: | 28 pages, 8 figures, 8 tables. Companion to arXiv:2610.16043. Code, paper source and figures: this https URL models and corpus: this https URL |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.37484 [cs.CL] |
| (or arXiv:2610.37484v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.37484 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Trent Northen [view email]
[v1]
Tue, 27 Oct 2026 16:11:04 UTC (192 KB)
来源：arXiv cs.CL · arxiv.org