🛰️ Daily AI Frontier
‹ back to 2026-09-07

Forgetting Without Restarting: Execution-State Unlearning for Stateful LLM Agents

Research LLM Agents

Ranking

Overall 81
Content 100
Popularity 37

Observed public metrics from 1 member.

Representative image for Forgetting Without Restarting: Execution-State Unlearning for Stateful LLM Agents

Merged summary

TL;DR - This paper formalizes execution-state unlearning for stateful LLM agents and proposes provenance-guided selective replay to remove revoked information from prompts, memory, plans, and KV caches. The method matches a full reset’s behavior while recomputing up to 9× fewer tokens.

  • Exact unlearning requires recomputing at least (T-\tau+1) transitions after the target information’s injection step.
  • Selective replay restores a pre-injection checkpoint, crops the KV cache, and regenerates the affected suffix from sanitized state.
  • Deleting memory records alone did not reduce leakage, while instruction-based forgetting reached Leak@probes = 1.00 under elicitation.
  • Source redaction still applied a revoked preference in 80% of episodes; selective replay was indistinguishable from a full reset in the reported audits.

Sources (1)

Forgetting Without Restarting: Execution-State Unlearning for Stateful LLM Agents

arXiv cs.CR Chao Yao, Yangbo Wei, Zhen Huang, Junhong Qian, Chenle Chen, Shaoqiang Lu, Chen Wu, Lei He 2026-09-04 arXiv:2609.04875
Public signals Semantic Scholar citations 0 · Semantic Scholar influential citations 0
Providers: Hugging Face · N/A OpenAlex · N/A Publisher · N/A Semantic Scholar · Citations 0 · Influential citations 0 X · N/A Fetched 2026-09-17 14:22:57.943057 UTC

TL;DR - This paper formalizes execution-state unlearning for stateful LLM agents and proposes provenance-guided selective replay to remove revoked information from prompts, memory, plans, and KV caches. The method matches a full reset’s behavior while recomputing up to 9× fewer tokens.

  • Exact unlearning requires recomputing at least (T-\tau+1) transitions after the target information’s injection step.
  • Selective replay restores a pre-injection checkpoint, crops the KV cache, and regenerates the affected suffix from sanitized state.
  • Deleting memory records alone did not reduce leakage, while instruction-based forgetting reached Leak@probes = 1.00 under elicitation.
  • Source redaction still applied a revoked preference in 80% of episodes; selective replay was indistinguishable from a full reset in the reported audits.
item →