Pith Number

pith:TATCZOFL

pith:2026:TATCZOFLYXKDUXOV6FHBF3UMGM

not attested not anchored not stored refs resolved

Dimension-Level Intent Fidelity Evaluation for Large Language Models: Evidence from Structured Prompt Ablation

Gang Peng

Many LLM outputs with perfect holistic scores still miss user intent on specific dimensions.

arxiv:2605.14517 v1 · 2026-05-14 · 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{TATCZOFLYXKDUXOV6FHBF3UMGM}

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

among Chinese-language outputs with complete paired scores, 25.7% received perfect holistic alignment scores (GA=5) while exhibiting measurable dimensional intent deficits; among English-language outputs, this proportion rose to 58.6%.

C2weakest assumption

That the structured prompt ablation and proxy annotation reliably isolate prior inferability from default recoverability without introducing selection bias or confounding the human validation of split-zone outputs.

C3one line summary

Dimension-level evaluation reveals that 25-58% of LLM outputs with perfect holistic scores still show measurable intent deficits across languages and domains.

References

18 extracted · 18 resolved · 4 Pith anchors

[1] Huang, L. et al. A survey on hallucination in large language models. ACM Comput. Surv. 57, 1–38 (2023) 2023

[2] Ji, Z. et al. Survey of hallucination in natural language generation. ACM Comput. Surv. 55, 1–38 (2023) 2023

[3] A Survey of Hallucination in Large Foundation Models 2023 · arXiv:2309.05922

[4] Lewis, P. et al. Retrieval -augmented generation for knowledge -intensive NLP tasks. NeurIPS (2020) 2020

[5] Ouyang, L. et al. Training language models to follow instructions with human feed back. NeurIPS (2022) 2022

Receipt and verification

First computed	2026-05-17T23:39:06.111116Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

98262cb8abc5d43a5dd5f14e12ee8c332f7f93f4288ae09227c2c39e386d8774

Aliases

arxiv: 2605.14517 · arxiv_version: 2605.14517v1 · doi: 10.48550/arxiv.2605.14517 · pith_short_12: TATCZOFLYXKD · pith_short_16: TATCZOFLYXKDUXOV · pith_short_8: TATCZOFL

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/TATCZOFLYXKDUXOV6FHBF3UMGM \
  | 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: 98262cb8abc5d43a5dd5f14e12ee8c332f7f93f4288ae09227c2c39e386d8774

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "19e0caffa6a0c51973879c6faeeb44b896269924f87986fbb66398bca62d093f",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.CL",
    "submitted_at": "2026-05-14T08:00:23Z",
    "title_canon_sha256": "5ac0bc7d8a0afe706e34908ca025e05f0c65720dad7e87369c1b63f1d7f4d1a0"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.14517",
    "kind": "arxiv",
    "version": 1
  }
}