Skip to content
OpenTrain AIFor AI Companies
← Back to explorer

MemDelta: Controlled Baselines and Hidden Confounds in Agent Memory Evaluation

Kuan Wang · Jun 29, 2026 · Citations: 0

Automatic Metrics
Open arXiv Find Implementation RSS feed Shortlist (0)

How to use this page

Low trust

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

Best use

Background context only

What to verify

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

Evidence quality

Low

Derived from extracted protocol signals and abstract evidence.

Abstract

Agent memory systems are increasingly evaluated against RAG and full-context baselines, but reported gains often mix changes in the memory method with changes in the language model, embedding model, or retrieval pipeline, making it unclear what is actually being measured. We present MemDelta, a controlled evaluation protocol that varies one component at a time on LongMemEval-S (500 questions, 50+ sessions, three model families). Four findings emerge: (1) verbatim RAG matches full-context GPT-4o-mini (47.2% vs. 49.8%, p = 0.34), but the ranking reverses across models: Gemini gains +14pp from full context, while Sonnet gains +31pp from RAG, partly because it refuses 63% of full-context queries; (2) swapping only the embedding model in an identical pipeline shifts accuracy by +6.2pp at n = 500 (p = 0.004), and Mem0 beats MiniLM-RAG by +11pp but loses to cloud-RAG by 1.2pp, so one variable flips the conclusion; (3) agent self-memory (42%) underperforms basic retrieval (47%); (4) on 2 of 6 question types (n = 88), Mem0 matches cloud RAG (72.7% vs. 73.9%, p = 1.0) at 50x the cost, suggesting narrow rather than general gains. We recommend memory evaluations fix embedding models across comparisons, stratify by model family, and report write-path cost before attributing gains to architecture.

Abstract-only analysis — low confidence

All signals on this page are inferred from the abstract only and may be inaccurate. Do not use this page as a primary protocol reference.

  • This paper looks adjacent to evaluation work, but not like a strong protocol reference.
  • The available metadata is too thin to trust this as a primary source.
Human Eval Hub LLM-as-Judge Hub Pairwise Preference Hub

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.

Best use

Background context only

Use if you need

A benchmark-and-metrics comparison anchor.

Main weakness

This paper looks adjacent to evaluation work, but not like a strong protocol reference.

Trust level

Low

Usefulness score

5/100 • Low

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

Human Feedback Signal

Not explicit in abstract metadata

Evaluation Signal

Detected

Usefulness for eval research

Adjacent candidate

Extraction confidence 45%

If you are doing eval pipeline work, start here:

Human Eval Hub LLM-as-Judge Hub Pairwise Preference Hub Tool-Use Eval Hub

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.

"Agent memory systems are increasingly evaluated against RAG and full-context baselines, but reported gains often mix changes in the memory method with changes in the language model, embedding model, or retrieval pipeline, making it unclear what is actually being measured."

Evaluation Modes

partial

Automatic Metrics

Includes extracted eval setup.

"Agent memory systems are increasingly evaluated against RAG and full-context baselines, but reported gains often mix changes in the memory method with changes in the language model, embedding model, or retrieval pipeline, making it unclear what is actually being measured."

Quality Controls

missing

Not reported

No explicit QC controls found.

"Agent memory systems are increasingly evaluated against RAG and full-context baselines, but reported gains often mix changes in the memory method with changes in the language model, embedding model, or retrieval pipeline, making it unclear what is actually being measured."

Benchmarks / Datasets

partial

Longmemeval

Useful for quick benchmark comparison.

"We present MemDelta, a controlled evaluation protocol that varies one component at a time on LongMemEval-S (500 questions, 50+ sessions, three model families)."

Reported Metrics

partial

Accuracy

Useful for evaluation criteria comparison.

"49.8%, p = 0.34), but the ranking reverses across models: Gemini gains +14pp from full context, while Sonnet gains +31pp from RAG, partly because it refuses 63% of full-context queries; (2) swapping only the embedding model in an identical pipeline shifts accuracy by +6.2pp at n = 500 (p = 0.004), and Mem0 beats MiniLM-RAG by +11pp but loses to cloud-RAG by 1.2pp, so one variable flips the conclusion; (3) agent self-memory (42%) underperforms basic retrieval (47%); (4) on 2 of 6 question types (n = 88), Mem0 matches cloud RAG (72.7% vs."

Human Feedback Details

  • Uses human feedback: No
  • Feedback types: None
  • Rater population: Not reported
  • Unit of annotation: Ranking (inferred)
  • Expertise required: General

Evaluation Details

  • Evaluation modes: Automatic Metrics
  • Agentic eval: None
  • Quality controls: Not reported
  • Evidence quality: Low
  • Use this page as: Background context only

Protocol And Measurement Signals

Benchmarks / Datasets

Longmemeval

Reported Metrics

accuracy

Research Brief

Metadata summary

Agent memory systems are increasingly evaluated against RAG and full-context baselines, but reported gains often mix changes in the memory method with changes in the language model, embedding model, or retrieval pipeline, making it unclear what is actually being measured.

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

Key Takeaways

  • Agent memory systems are increasingly evaluated against RAG and full-context baselines, but reported gains often mix changes in the memory method with changes in the language model, embedding model, or retrieval pipeline, making it unclear what is actually being measured.
  • We present MemDelta, a controlled evaluation protocol that varies one component at a time on LongMemEval-S (500 questions, 50+ sessions, three model families).
  • Four findings emerge: (1) verbatim RAG matches full-context GPT-4o-mini (47.2% vs.

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

Research Summary

Contribution Summary

  • Agent memory systems are increasingly evaluated against RAG and full-context baselines, but reported gains often mix changes in the memory method with changes in the language model, embedding model, or retrieval pipeline, making it unclear…
  • We present MemDelta, a controlled evaluation protocol that varies one component at a time on LongMemEval-S (500 questions, 50+ sessions, three model families).
  • 49.8%, p = 0.34), but the ranking reverses across models: Gemini gains +14pp from full context, while Sonnet gains +31pp from RAG, partly because it refuses 63% of full-context queries; (2) swapping only the embedding model in an identical…

Why It Matters For Eval

  • We present MemDelta, a controlled evaluation protocol that varies one component at a time on LongMemEval-S (500 questions, 50+ sessions, three model families).
  • 49.8%, p = 0.34), but the ranking reverses across models: Gemini gains +14pp from full context, while Sonnet gains +31pp from RAG, partly because it refuses 63% of full-context queries; (2) swapping only the embedding model in an identical…

Researcher Checklist

  • Gap: Human feedback protocol is explicit

    No explicit human feedback protocol detected.

  • Pass: Evaluation mode is explicit

    Detected: Automatic Metrics

  • Gap: Quality control reporting appears

    No calibration/adjudication/IAA control explicitly detected.

  • Pass: Benchmark or dataset anchors are present

    Detected: Longmemeval

  • Pass: Metric reporting is present

    Detected: accuracy

Related Papers

Papers are ranked by protocol overlap, extraction signal alignment, and semantic proximity.