Pith Number

pith:NWZ2LUZQ

pith:2022:NWZ2LUZQPZPVMV62NUPT4I7GKI

not attested not anchored not stored refs pending

CodeGen: An Open Large Language Model for Code with Multi-Turn Program Synthesis

Bo Pang, Caiming Xiong, Erik Nijkamp, Hiroaki Hayashi, Huan Wang, Lifu Tu, Silvio Savarese, Yingbo Zhou

The same problem intent given to CodeGen in multiple turns produces substantially better programs than a single turn on the MTPB benchmark.

arxiv:2203.13474 v5 · 2022-03-25 · cs.LG · cs.CL · cs.PL

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{NWZ2LUZQPZPVMV62NUPT4I7GKI}

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

Our analysis on MTPB shows that the same intent provided to CODEGEN in multi-turn fashion significantly improves program synthesis over that provided as a single turn.

C2weakest assumption

The multi-turn prompts in MTPB accurately capture natural human problem decomposition without introducing selection bias or making the benchmark easier than real-world use cases.

C3one line summary

CodeGen models match prior SOTA on HumanEval zero-shot code generation, and multi-turn prompting on the new MTPB benchmark significantly improves synthesis over single-turn inputs.

Cited by

60 papers in Pith

Rethinking Code Performance Benchmarks for LLMs

Accuracy and Satisfaction in Multi-Turn LLM Dialogues for NFR Assessment

An Empirical Study of LLM-Generated Specifications for VeriFast

CodeTeam: An LLM-Powered Multi-Agent Framework for Repository-Level Code Generation

Multi-LCB: Extending LiveCodeBench to Multiple Programming Languages

Receipt and verification

First computed	2026-07-05T05:46:14.025352Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

6db3a5d3307e5f5657da6d1f3e23e65215f3144e17649beb2da19056b17d8538

Aliases

arxiv: 2203.13474 · arxiv_version: 2203.13474v5 · doi: 10.48550/arxiv.2203.13474 · pith_short_12: NWZ2LUZQPZPV · pith_short_16: NWZ2LUZQPZPVMV62 · pith_short_8: NWZ2LUZQ

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/NWZ2LUZQPZPVMV62NUPT4I7GKI \
  | 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: 6db3a5d3307e5f5657da6d1f3e23e65215f3144e17649beb2da19056b17d8538

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "d936db27aa7f28c7873b54f880d9ed3a292e1595afe3a75199ea36c7bd3f6dab",
    "cross_cats_sorted": [
      "cs.CL",
      "cs.PL"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2022-03-25T06:55:15Z",
    "title_canon_sha256": "9ba27ade9785782edaa81753e0dc8d1830e1e65541a9e8a1d83f1c4511ff9249"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2203.13474",
    "kind": "arxiv",
    "version": 5
  }
}