Oriel — AI & technology worth noticing, every day.

Zero-Mem: Zero-Token Memory Operations for LLM Agents

Paper · https://arxiv.org/abs/2607.29377

LLM agents need memory to act consistently over long interactions, yet many systems use additional LLM calls to operate that memory. Generating intermediate records and mediating their retrieval adds recurring token and time costs, while omitted or merged details can obscure the original evidence. We ask whether structured memory access requires generation at all. Zero-Mem introduces zero-token memory operations: no step outside final question answering invokes an LLM or consumes LLM input or output tokens; encoder computation is accounted for separately. Zero-Mem preserves original interaction traces as its source of record. It organizes the traces in two complementary ways. An entity--context graph exposes connections across interactions, while a temporal hierarchy preserves conversational locality and session state. For each query, Zero-Mem weighs the two views, retrieves from both, and follows their structure to recover supporting relations or surrounding context. Deterministic calibration first discards conflicting evidence and then keeps the reader's answer grounded in the retrieved traces. Only the final-QA reader invokes an LLM. Across long-memory and long-context question-answering benchmarks, Zero-Mem achieves competitive performance while eliminating LLM calls and LLM-token consumption from memory operations. With the same final-QA reader and context budget, it reduces memory-operation time cost by 57.6\% relative to the fastest compared baseline. Ablations support the contribution of the two views and their query-dependent coordination. Overall, the results show that structured agent memory need not generate an intermediate representation of the past. After peer review, the code and implementation details will be available at blue{https://github.com/TheMoon0815/Zero-mem}.

First seen by Oriel on .

AI summary

Zero-Mem stores and retrieves LLM agent memory without LLM calls or token use during memory operations, while achieving competitive question-answering performance.

  • It preserves original interaction traces and organizes them with an entity-context graph and a temporal hierarchy for conversational locality and session state.
  • For each query, it retrieves from both structures, weights their contributions, and uses deterministic calibration to discard conflicting evidence and ground answers in traces.
  • Only the final question-answering reader invokes an LLM; encoder computation is accounted for separately.
  • With the same reader and context budget, Zero-Mem reduces memory-operation time by 57.6% relative to the fastest compared baseline.
  • Ablations support the value of both memory views and their query-dependent coordination; code and implementation details are planned after peer review.

#llm agents #agent memory #question answering

Generated by gpt-5.6-sol. Check the original for details.

Metric trends

Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained

Points

Score reported by the source.

  • Hugging Face Papers: 13, latest observation ; 0 over observed period since

Comments

Comment count reported by the source.

  • Hugging Face Papers: 2, latest observation ; 0 over observed period since

Aliases

SourceExternal IDURL
Hugging Face Papers2607.29377https://arxiv.org/abs/2607.29377 Read discussion