Pith Number

pith:LCB3UHAA

pith:2025:LCB3UHAAN6GMCNM5457GPWNF6W

not attested not anchored not stored refs pending

VL-Rethinker: Incentivizing Self-Reflection of Vision-Language Models with Reinforcement Learning

Chao Qu, Fangzhen Lin, Haozhe Wang, Wei Chu, Wenhu Chen, Zuming Huang

Reinforcement learning with selective replay and forced rethinking steps lets vision-language models reflect on their answers and reach new highs on multimodal math benchmarks.

arxiv:2504.08837 v3 · 2025-04-10 · 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{LCB3UHAAN6GMCNM5457GPWNF6W}

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

By combining Selective Sample Replay and Forced Rethinking in RL training, VL-Rethinker advances state-of-the-art scores on MathVista to 80.4% and MathVerse to 63.5%, achieving open-source SoTA on MathVision, MMMU-Pro, EMMA, and MEGA-Bench.

C2weakest assumption

That the reported gains stem primarily from increased self-reflection and slow-thinking rather than from other side effects of the RL setup or from benchmark-specific optimizations.

C3one line summary

VL-Rethinker reaches 80.4% on MathVista and 63.5% on MathVerse by adapting GRPO with Selective Sample Replay and Forced Rethinking to promote self-reflection in vision-language models without distillation.

Formal links

2 machine-checked theorem links

Cited by

50 papers in Pith

Learning Spatiotemporal Sensitivity in Video LLMs via Counterfactual Reinforcement Learning

Visual-Advantage On-Policy Distillation for Vision-Language Models

Are VLMs Seeing or Just Saying? Uncovering the Illusion of Visual Re-examination

GRASP: Learning to Ground Social Reasoning in Multi-Person Non-Verbal Interactions

Reasoning Portability: Guiding Continual Learning for MLLMs in the RLVR Era

Receipt and verification

First computed	2026-05-17T23:38:53.290891Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

5883ba1c006f8cc1359de77e67d9a5f5b64427899b130d9f6a57ce29a199434e

Aliases

arxiv: 2504.08837 · arxiv_version: 2504.08837v3 · doi: 10.48550/arxiv.2504.08837 · pith_short_12: LCB3UHAAN6GM · pith_short_16: LCB3UHAAN6GMCNM5 · pith_short_8: LCB3UHAA

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/LCB3UHAAN6GMCNM5457GPWNF6W \
  | 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: 5883ba1c006f8cc1359de77e67d9a5f5b64427899b130d9f6a57ce29a199434e

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "c32ff01e1d574c8407f7d747be89e3134028b09fc84bdb3106f98de9fc65a742",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2025-04-10T17:41:56Z",
    "title_canon_sha256": "44ab9014d6ac1819ca9747e6da46d768642c25e8940acc1356b7d8608ad8420c"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2504.08837",
    "kind": "arxiv",
    "version": 3
  }
}