pith:J7F5OYQ5
Epistemic Regret Minimization: Label-Free Causal Critique Beyond Outcome Reward
Epistemic Regret Minimization identifies causal flaws in LLM reasoning traces without ground-truth labels
arxiv:2602.11675 v4 · 2026-02-12 · cs.AI
Add to your LaTeX paper
\usepackage{pith}
\pithnumber{J7F5OYQ5MHRNOYL3PJQ762N64A}
Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge
Record completeness
Claims
A separation theorem proves outcome-only RL cannot distinguish correct from flawed causal models in confounded environments, and preliminary experiments show epistemic reward carries signal where outcome reward does not.
That causal flaws are reliably identifiable and correctable from reasoning traces alone without ground-truth labels or external verifiers, and that the observed corrections on CausalT5K and CLadder generalize beyond the tested models and scenarios.
Epistemic Regret Minimization identifies causal reasoning flaws in LLMs from traces alone, corrects stubborn models where outcome-only methods fail, and is supported by a separation theorem proving outcome-only RL cannot distinguish correct from flawed causal models in confounded settings.
Formal links
Receipt and verification
| First computed | 2026-05-21T01:04:23.151766Z |
|---|---|
| Builder | pith-number-builder-2026-05-17-v1 |
| Signature | Pith Ed25519
(pith-v1-2026-05) · public key |
| Schema | pith-number/v1.0 |
Canonical hash
4fcbd7621d61e2d7617b7a61ff69bee007aae4e7dab92eec6ed8ab2c6bf44248
Aliases
· · · · ·Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/J7F5OYQ5MHRNOYL3PJQ762N64A \
| 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: 4fcbd7621d61e2d7617b7a61ff69bee007aae4e7dab92eec6ed8ab2c6bf44248
Canonical record JSON
{
"metadata": {
"abstract_canon_sha256": "661e71f4dec067c55314260d28df484caf64fb8bcb838baa0a4d0a30ba4890d9",
"cross_cats_sorted": [],
"license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
"primary_cat": "cs.AI",
"submitted_at": "2026-02-12T07:48:21Z",
"title_canon_sha256": "dfc6af3c865c59bb01d646c95472519247da114c9af32f42b085052598999647"
},
"schema_version": "1.0",
"source": {
"id": "2602.11675",
"kind": "arxiv",
"version": 4
}
}