Pith Number

pith:2GWQW6DN

pith:2026:2GWQW6DNCH3NVSCZ32X5OPYG6X

not attested not anchored not stored refs pending

SUPERNOVA: Eliciting General Reasoning in LLMs with Reinforcement Learning on Natural Instructions

Ashima Suvarna, Hritik Bansal, Kendrick Phan, Mehrab Beikzadeh, Saadia Gabriel

Curating expert-annotated instruction data for verifiable rewards extends reinforcement learning to general reasoning tasks in language models.

arxiv:2604.08477 v2 · 2026-04-09 · cs.AI · cs.CL · cs.LG

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{2GWQW6DNCH3NVSCZ32X5OPYG6X}

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

models trained on SUPERNOVA outperform strong baselines (e.g., Qwen3.5) on challenging reasoning benchmarks including BBEH, Zebralogic, and MMLU-Pro. In particular, training on SUPERNOVA yields relative improvements of up to 52.8% on BBEH across model sizes, demonstrating the effectiveness of principled data curation for RLVR.

C2weakest assumption

That instruction-tuning datasets with expert-annotated ground-truth encode rich reasoning patterns that can be systematically adapted into high-quality verifiable rewards for RLVR, and that the 100+ controlled experiments isolate the effects of source task selection and mixing from other training variables.

C3one line summary

SUPERNOVA adapts instruction-tuning data for RLVR and achieves up to 52.8% relative gains on general reasoning benchmarks like BBEH through targeted task selection and mixing.

Receipt and verification

First computed	2026-06-05T01:14:38.136786Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

d1ad0b786d11f6dac859deafd73f06f5fb63f710f3ada59bf4149f5d70d8a32a

Aliases

arxiv: 2604.08477 · arxiv_version: 2604.08477v2 · doi: 10.48550/arxiv.2604.08477 · pith_short_12: 2GWQW6DNCH3N · pith_short_16: 2GWQW6DNCH3NVSCZ · pith_short_8: 2GWQW6DN

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/2GWQW6DNCH3NVSCZ32X5OPYG6X \
  | 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: d1ad0b786d11f6dac859deafd73f06f5fb63f710f3ada59bf4149f5d70d8a32a

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "b48f6c434d015ab9c3cb1ea13298d5da80cbbd397cb1b2642cf9c656b42ebb08",
    "cross_cats_sorted": [
      "cs.CL",
      "cs.LG"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.AI",
    "submitted_at": "2026-04-09T17:16:07Z",
    "title_canon_sha256": "042144a56442920a5a3a992e56722ed146e6cde347139fb5d1b309a6e2222d0c"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2604.08477",
    "kind": "arxiv",
    "version": 2
  }
}