Persistent self-modifying AI agents exhibit compositional drift from mismatches across five mutability layers, with governance difficulty rising under rapid mutation, strong coupling, weak reversibility, and low observability, as indicated by a 0.68 identity hysteresis ratio in a preliminary ratchet
ReAct: Synergizing reasoning and acting in language models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.AI 1years
2026 1verdicts
UNVERDICTED 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Layered Mutability: Continuity and Governance in Persistent Self-Modifying Agents
Persistent self-modifying AI agents exhibit compositional drift from mismatches across five mutability layers, with governance difficulty rising under rapid mutation, strong coupling, weak reversibility, and low observability, as indicated by a 0.68 identity hysteresis ratio in a preliminary ratchet