06RESEARCH PAPERPreprint
Risk, Reward, and Survival Pressure in Artificial Agents
發表日期
摘要
ABSTRACTZenodo 原文Risk in SPVE is not only reward seeking. The same lobster can chase payout, avoid loss, react to Tiamat pressure, respond to social pressure, and enter repair. This draft separates reward, settlement, and survival pressure so the dashboard can answer a practical question: is the agent making a profitable decision, a fearful decision, a social decision, or a repair decision?
全文 PDF
FULL TEXT1 個檔- 下載 PDFpaper.pdf・50 KB・本站
PDF 與 Zenodo 上的檔案相同（大小與 md5 已核對）。
論文資料
DETAILSZenodo 紀錄 20743920- 作者
- Chen, Ho Yiing
- 發表日期
- 資源類型
- Preprint
- DOI
- 10.5281/zenodo.20743920
- Zenodo
- zenodo.org/records/20743920
- 授權
- Creative Commons Attribution 4.0 International
作者 Chen, Ho Yiing 是織田紀香的本名（陳禾穎）英文拼寫。標題、摘要、關鍵字與授權照 Zenodo 原文。
關鍵字
KEYWORDS24 個- SPVE
- AI agents
- persistent agents
- longitudinal agents
- multi-agent systems
- agent ecology
- agent observability
- lobster agents
- behavioral telemetry
- agent monitoring
- workflow ecology
- agent evaluation
- agent risk
- reward drive
- survival pressure
- betting agents
- trading agents
- raid settlements
- sports betting
- stock trades
- open positions
- realized P&L
- crisis avoidance
- risk ledger