pith:E6XBFBWF
Embodied Interpretability: Linking Causal Understanding to Generalization in Vision-Language-Action Models
Interventional attribution reveals when vision-language-action models depend on spurious features rather than true causes.
arxiv:2605.00321 v2 · 2026-05-01 · cs.RO
Add to your LaTeX paper
\usepackage{pith}
\pithnumber{E6XBFBWFCQUWDNKRCVRJAMMBBD}
Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge
Record completeness
Claims
Experiments across diverse manipulation tasks indicate that NMR predicts generalization behavior and that ISS yields more faithful explanations than existing interpretability methods.
That the interventional masking procedure yields unbiased estimates of causal influence on actions and that action prediction error serves as a valid proxy for causal influence under the characterized conditions.
Interventional attribution via ISS and NMR diagnoses causal misalignment in VLA policies and predicts their generalization performance across manipulation tasks.
Cited by
Receipt and verification
| First computed | 2026-06-11T01:10:36.917482Z |
|---|---|
| Builder | pith-number-builder-2026-05-17-v1 |
| Signature | Pith Ed25519
(pith-v1-2026-05) · public key |
| Schema | pith-number/v1.0 |
Canonical hash
27ae1286c5142961b551156290318108dfeab7749a840984e81ec33ff8f968f8
Aliases
· · · · ·Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/E6XBFBWFCQUWDNKRCVRJAMMBBD \
| 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: 27ae1286c5142961b551156290318108dfeab7749a840984e81ec33ff8f968f8
Canonical record JSON
{
"metadata": {
"abstract_canon_sha256": "dcbb8f8b88150556906fa30b4c98f9b6f363f884cd4ac17df6ff9877dbbcbdee",
"cross_cats_sorted": [],
"license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
"primary_cat": "cs.RO",
"submitted_at": "2026-05-01T01:00:00Z",
"title_canon_sha256": "080ec0cf66f20199d1138b95e1bb171f2aa0d7f25eb19d8f5c889318352be54e"
},
"schema_version": "1.0",
"source": {
"id": "2605.00321",
"kind": "arxiv",
"version": 2
}
}