Pith Number

pith:PCNMERUM

pith:2026:PCNMERUM4KEKZOQI4VMJ7NJHOS

not attested not anchored not stored refs pending

OGPO: Sample Efficient Full-Finetuning of Generative Control Policies

Abhishek Gupta, Chaoyi Pan, Cleah Winston, Douglas Chen, Giri Anantharaman, Hongkai Dai, Jesse Zhang, Manan Agarwal, Max Simchowitz, Mitsuhiko Nakamoto, Nai-Chieh Huang, Oliver Kroemer, Paarth Shah, Sarvesh Patil, Sergey Levine, Shashwat Saxena, Zeynep Temel

OGPO fine-tunes generative control policies to near full task success from poor initializations with no expert data.

arxiv:2605.03065 v2 · 2026-05-04 · cs.LG · cs.RO

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{PCNMERUM4KEKZOQI4VMJ7NJHOS}

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

OGPO achieves state-of-the-art performance on manipulation tasks spanning multi-task settings, high-precision insertion, and dexterous control. It is the only method that can fine-tune poorly-initialized behavior cloning policies to near full task-success with no expert data in the online replay buffer, and does so with few task-specific hyperparameter tuning.

C2weakest assumption

That off-policy critics can be maintained stably while propagating gradients through the full generative process of the policy via a modified PPO objective without introducing instability or requiring heavy per-task tuning across state- and pixel-based settings.

C3one line summary

OGPO is a sample-efficient off-policy method for full finetuning of generative control policies that reaches SOTA on robotic manipulation tasks and can recover from poor behavior-cloning initializations without expert data.

Formal links

2 machine-checked theorem links

Receipt and verification

First computed	2026-06-11T01:09:36.934198Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

789ac2468ce288acba08e5589fb527749ebe5668847bb47664ccf41ac702ca3a

Aliases

arxiv: 2605.03065 · arxiv_version: 2605.03065v2 · doi: 10.48550/arxiv.2605.03065 · pith_short_12: PCNMERUM4KEK · pith_short_16: PCNMERUM4KEKZOQI · pith_short_8: PCNMERUM

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/PCNMERUM4KEKZOQI4VMJ7NJHOS \
  | 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: 789ac2468ce288acba08e5589fb527749ebe5668847bb47664ccf41ac702ca3a

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "3a1340d177da08657e0d0a1971623f52c995294951d18469f33b5163c4185185",
    "cross_cats_sorted": [
      "cs.RO"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2026-05-04T18:36:40Z",
    "title_canon_sha256": "8e90aa034719a618774f63ed25aaa69d7d1dba226830d5d96d7cdf3d53126fe5"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.03065",
    "kind": "arxiv",
    "version": 2
  }
}