Pith Number

pith:PVG73FUF

pith:2026:PVG73FUFVM2W5O5TAGTS457QF3

not attested not anchored not stored refs pending

Spurious Correlation Learning in Preference Optimization: Mechanisms, Consequences, and Mitigation via Tie Training

Alex Semendinger, Christian Moya, Elliott Thornley, Guang Lin

Preference optimization induces spurious feature reliance through mean bias and correlation leakage, creating a vulnerability to distribution shift that more training data cannot fix.

arxiv:2605.11134 v2 · 2026-05-11 · cs.LG · 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{PVG73FUFVM2W5O5TAGTS457QF3}

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

standard preference-learning objectives induce reliance on spurious features at the population level through two channels: mean spurious bias and causal--spurious correlation leakage. We then show that this reliance creates an irreducible vulnerability to distribution shift: more data from the same training distribution fails to reduce the model's dependence on spurious features.

C2weakest assumption

The mechanisms and mitigation identified for log-linear policies extend to neural networks and large language models without significant degradation of causal learning.

C3one line summary

Standard preference learning induces spurious feature reliance via mean bias and correlation leakage, creating irreducible distribution shift vulnerabilities that tie training mitigates without degrading causal learning.

Cited by

1 paper in Pith

Out-of-Distribution Generalization of Risk Aversion in Language Models

Receipt and verification

First computed	2026-06-01T02:03:43.357095Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

7d4dfd9685ab356ebbb301a72e77f02efc669222652cd3613718eb87310fec74

Aliases

arxiv: 2605.11134 · arxiv_version: 2605.11134v2 · doi: 10.48550/arxiv.2605.11134 · pith_short_12: PVG73FUFVM2W · pith_short_16: PVG73FUFVM2W5O5T · pith_short_8: PVG73FUF

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/PVG73FUFVM2W5O5TAGTS457QF3 \
  | 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: 7d4dfd9685ab356ebbb301a72e77f02efc669222652cd3613718eb87310fec74

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "7cf849e07bfa31690c04496f70cfdcc5ef97ce2c7becb29977e80196f161735f",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://creativecommons.org/licenses/by-nc-nd/4.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2026-05-11T18:41:12Z",
    "title_canon_sha256": "64f159179108eb2e6201f854d383fb9a0adf09c38557b327780868ad7c36db05"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.11134",
    "kind": "arxiv",
    "version": 2
  }
}