Pith Number

pith:L2KZ6GFW

pith:2026:L2KZ6GFWAF2IBCT3ZJGMTHYDRF

not attested not anchored not stored refs resolved

RTLC -- Research, Teach-to-Learn, Critique: A three-stage prompting paradigm inspired by the Feynman Learning Technique that lifts LLM-as-judge accuracy on JudgeBench with no fine-tuning

Andrea Morandi

A three-stage prompting method lifts LLM judge accuracy from 65% to 79% on hard pairwise comparisons.

arxiv:2605.13695 v1 · 2026-05-13 · cs.CL · cs.AI

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{L2KZ6GFWAF2IBCT3ZJGMTHYDRF}

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

On JudgeBench-GPT (350 hard pairwise items), Claude 3.7 Sonnet's pairwise accuracy climbs from 64.6% (single-shot vanilla prompt) to 78.6% (RTLC critique-of-10) -- an absolute 14.0-percentage-point gain.

C2weakest assumption

That the reported accuracy gains are driven by the specific RTLC stages rather than increased token budget, model-specific behavior, or benchmark idiosyncrasies, and that the high-level stage descriptions translate to reproducible prompts.

C3one line summary

RTLC prompting lifts Claude 3.7 Sonnet pairwise accuracy on 350 hard JudgeBench items from 64.6% to 78.6% via a Research-Teach-Critique scaffold that beats self-consistency.

References

10 extracted · 10 resolved · 5 Pith anchors

[1] JudgeBench: A Benchmark for Evaluating LLM-based Judges 2025 · arXiv:2410.12784

[2] Two ways to de-bias an LLM-as-a-Judge: A continuous- score comparison of hierarchical Bayesian calibration and Neural-ODE score transport, 2026

[3] Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena 2023

[4] UltraFeedback: Boosting language models with scaled AI feedback, 2024

[5] Self-Consistency Improves Chain of Thought Reasoning in Language Models 2023 · arXiv:2203.11171

Cited by

1 paper in Pith

Sequential Consensus for Multi-Agent LLM Debates: A Wald-SPRT compute governor with calibration-based failure detection

Receipt and verification

First computed	2026-05-18T02:44:16.910577Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

5e959f18b60174808a7bca4cc99f038947fd79f6f4dd028bd38b851e3f8fb831

Aliases

arxiv: 2605.13695 · arxiv_version: 2605.13695v1 · doi: 10.48550/arxiv.2605.13695 · pith_short_12: L2KZ6GFWAF2I · pith_short_16: L2KZ6GFWAF2IBCT3 · pith_short_8: L2KZ6GFW

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/L2KZ6GFWAF2IBCT3ZJGMTHYDRF \
  | 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: 5e959f18b60174808a7bca4cc99f038947fd79f6f4dd028bd38b851e3f8fb831

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "a356867732d5cd804f47160ad9c6aef81f170696de37075918f5dfdbe0367e1c",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.CL",
    "submitted_at": "2026-05-13T15:48:16Z",
    "title_canon_sha256": "c4779848ed5081b59722c936a564a648a6a165e56430d251b11740f808565f23"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.13695",
    "kind": "arxiv",
    "version": 1
  }
}