Pith Number

pith:QZVCWZPC

pith:2026:QZVCWZPCXSQYQRXNNRUXFWCNGR

not attested not anchored not stored refs resolved

R^2-Mem: Reflective Experience for Memory Search

Junkang Wu, Wenyu Mao, Xiangnan He, Xiang Wang, Xinyuan Wang

R^2-Mem distills abstract experiences from scored past trajectories to guide LLM agents away from repeated search errors without reinforcement learning.

arxiv:2605.13486 v1 · 2026-05-13 · cs.CL

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{QZVCWZPCXSQYQRXNNRUXFWCNGR}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

R^2-Mem provides a RL-free and low-cost solution for self-improving LLM agents, improving F1 scores by up to 22.6%, while reducing token consumption by 12.9% and search iterations by 20.2%.

C2weakest assumption

That the rubric-guided evaluator accurately distinguishes high- and low-quality steps and that the distilled experiences generalize to new queries without overfitting to the offline trajectories.

C3one line summary

R^2-Mem distills rubric-scored experiences from high- and low-quality search trajectories to guide LLM agents, raising F1 by up to 22.6% while cutting tokens 12.9% and iterations 20.2%.

References

11 extracted · 11 resolved · 1 Pith anchors

[1] M 3-embedding: Multi-linguality, multi-functionality, multi-granularity text embeddings through self-knowledge distillation 2024 · doi:10.18653/v1/2024.findings-acl.137

[2] The N arrative QA Reading Comprehension Challenge 2026 · doi:10.1162/tacl_a_00023

[3] URLhttps://openreview.net/forum?id=jL7fwchScm. William F. Shen, Xinchi Qiu, Chenxi Whitehouse, Lisa Alazraki, Shashwat Goel, Francesco Barbieri, Timon Willi, Akhil Mathur, and Ilias Leontiadis. Rethin 2026

[4] Memory-R1: Enhancing Large Language Model Agents to Manage and Utilize Memories via Reinforcement Learning 2018 · arXiv:2508.19828

[5] Analyze why this planning trace is high-quality or low-quality

Receipt and verification

First computed	2026-05-18T02:44:41.265847Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

866a2b65e2bca18846ed6c6972d84d344d8356b8094bb826ed3cdcdbff15f6ac

Aliases

arxiv: 2605.13486 · arxiv_version: 2605.13486v1 · doi: 10.48550/arxiv.2605.13486 · pith_short_12: QZVCWZPCXSQY · pith_short_16: QZVCWZPCXSQYQRXN · pith_short_8: QZVCWZPC

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/QZVCWZPCXSQYQRXNNRUXFWCNGR \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: 866a2b65e2bca18846ed6c6972d84d344d8356b8094bb826ed3cdcdbff15f6ac

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "44514237febfe7091fed30834c9849bf363845afb57451f69902fcfce6bf9715",
    "cross_cats_sorted": [],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.CL",
    "submitted_at": "2026-05-13T13:09:36Z",
    "title_canon_sha256": "649b4779aabb9fccb77f698a33dbc2290f0ffebd8a0bbe469d355409cdfa7fd3"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.13486",
    "kind": "arxiv",
    "version": 1
  }
}