Skip to content
OpenTrain AIFor AI Companies

HFEPX · Eval paper review

Persistent Memory in Multi-Agent LLM Inference: What It Costs, What It Buys, and When You Can Tell

Hochan Son, Kyungdoe Han, Jaehan Koh, Xiaowu Dai +2 more

Published

Oct 6, 2026

Citations

0

Trust level

Low

Usefulness score

25/100 (Low)

Extraction confidence

45% (Low)

Derived from extracted protocol signals and abstract evidence.

Rater population

Not reported

Signals refreshed

Oct 6, 2026

Open arXiv Find implementation Shortlist (0) RSS feed

Tags

Automatic Metrics General Multi Agent

Should you rely on this paper?

This paper is adjacent to HFEPX scope and is best used for background context, not as a primary protocol reference.

Use this as background context only. Do not make protocol decisions from this page alone.

Best use

Background context only

Use if you need

A secondary eval reference to pair with stronger protocol papers.

What to verify

Validate the evaluation procedure and quality controls in the full paper before operational use.

Main weakness

The available metadata is too thin to trust this as a primary source.

Human feedback signal
Not explicit
Not explicit in abstract metadata
Evaluation signal
Detected
Eval setup described
Usefulness for eval research
25/100
Adjacent candidate

Treat as adjacent context, not a core eval-method reference.

If you are doing eval pipeline work, start here

Human Eval Papers LLM-as-Judge Papers Pairwise Preference Papers Tool-Use Evaluation Search

Abstract

Decomposing long-context inference across cooperating agents bounds the active KV cache per call rather than total evidence, which matters when KV-cache memory binds. Many such systems add a persistent tier storing and recalling reasoning traces, usually validated by an ablation reporting an accuracy gain. We measure both on one three-tier agent architecture. Decomposition delivers: peak KV working set of 14.3 MiB per query against 35.5 and 35.3 MiB for single-pass and retrieval-augmented baselines. The persistent tier does not: across eight controlled dataset pairs at n=100 per arm it costs +0.368 MiB [+0.167, +0.590] of peak cache and produces no detectable accuracy change (+0.015, 95% CI [-0.011, +0.046]). We argue the null is structural: single-question benchmarks supply each item with its own evidence and score it independently, and correctness requires resetting stored traces between conditions, so recall has nothing informative to retrieve. Reaching it took four measurement corrections -- three inflating the apparent benefit, the fourth making an effect that size look resolvable -- none visible in the results table. We give the conditions an agent-memory ablation must satisfy and detection procedures that need no knowledge of the specific defect.

What we could verify

These are the protocol signals we could actually recover from the available paper metadata. Use them to decide whether this paper is worth deeper reading.

Human Feedback Types

missing

None explicit

No explicit feedback protocol extracted.

"Decomposing long-context inference across cooperating agents bounds the active KV cache per call rather than total evidence, which matters when KV-cache memory binds."

Evaluation Modes

partial

Automatic Metrics

Includes extracted eval setup.

"Decomposing long-context inference across cooperating agents bounds the active KV cache per call rather than total evidence, which matters when KV-cache memory binds."

Quality Controls

missing

Not reported

No explicit QC controls found.

"Decomposing long-context inference across cooperating agents bounds the active KV cache per call rather than total evidence, which matters when KV-cache memory binds."

Benchmarks / Datasets

missing

Not extracted

No benchmark anchors detected.

"Decomposing long-context inference across cooperating agents bounds the active KV cache per call rather than total evidence, which matters when KV-cache memory binds."

Reported Metrics

partial

Accuracy, Recall

Useful for evaluation criteria comparison.

"Many such systems add a persistent tier storing and recalling reasoning traces, usually validated by an ablation reporting an accuracy gain."

Benchmarks and datasets

No benchmark or dataset names were extracted from the available abstract.

Reported metrics

accuracyrecall
Human feedback details
Uses human feedback
No
Feedback types
None
Rater population
Not reported
Expertise required
General
Evaluation details
Evaluation modes
Automatic Metrics
Agentic eval
Multi Agent
Quality controls
Not reported
Evidence quality
Low
Use this page as
Background context only

Research brief

Metadata summary

Decomposing long-context inference across cooperating agents bounds the active KV cache per call rather than total evidence, which matters when KV-cache memory binds.

Based on abstract + metadata only. Check the source paper before making high-confidence protocol decisions.

Key takeaways

  • Decomposing long-context inference across cooperating agents bounds the active KV cache per call rather than total evidence, which matters when KV-cache memory binds.
  • Many such systems add a persistent tier storing and recalling reasoning traces, usually validated by an ablation reporting an accuracy gain.
  • We measure both on one three-tier agent architecture.

Researcher actions

  • Compare this paper against nearby papers in the same arXiv category before using it for protocol decisions.
  • Validate inferred eval signals (Automatic metrics) against the full paper.
  • Use related-paper links to find stronger protocol-specific references.

Caveats

  • Generated from abstract + metadata only; no PDF parsing.
  • Signals below are heuristic and may miss details reported outside the abstract.

Recommended queries

Accuracy

Contribution summary

  • Decomposing long-context inference across cooperating agents bounds the active KV cache per call rather than total evidence, which matters when KV-cache memory binds.
  • We measure both on one three-tier agent architecture.
  • We argue the null is structural: single-question benchmarks supply each item with its own evidence and score it independently, and correctness requires resetting stored traces between conditions, so recall has nothing informative to…

Why it matters for eval

  • Decomposing long-context inference across cooperating agents bounds the active KV cache per call rather than total evidence, which matters when KV-cache memory binds.
  • We measure both on one three-tier agent architecture.

Researcher checklist

  • Human feedback protocol is explicit

    No explicit human feedback protocol detected.

  • Evaluation mode is explicit

    Detected: Automatic Metrics

  • Quality control reporting appears

    No calibration/adjudication/IAA control explicitly detected.

  • Benchmark or dataset anchors are present

    No benchmark/dataset anchor extracted from abstract.

  • Metric reporting is present

    Detected: accuracy, recall