TasksSotADatasetsPapersMethodsSubmitAbout
Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable BenchmarksAll SotADatasetsPapersMethods

Community

Submit ResultsAbout

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Models/GPT-J 6B

GPT-J 6B

Reported on 9 benchmarks across 2 tasks · 2 papers

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing9 results

  • Code GenerationonVerified Smart Contract Code Comments
    BLEU score· 2023-09-18
    25.8
    best: 55.7 (GPT-J 6B Smart Contract)
    Efficient Avoidance of Vulnerabilities in Auto-completed Smart Contract Code Using Vulnerability-constrained DecodingarXiv:2309.09826
  • Question AnsweringonTruthfulQA
    % info· 2021-09-08
    89.96
    best: 97.7 (Alpaca 7B + Inference Time Intervention (ITI))
    TruthfulQA: Measuring How Models Mimic Human FalsehoodsarXiv:2109.07958
  • Question AnsweringonTruthfulQA
    % true· 2021-09-08
    26.68
    best: 88.6 (Vicuna 7B + Inference Time Intervention (ITI))
    TruthfulQA: Measuring How Models Mimic Human FalsehoodsarXiv:2109.07958
  • Question AnsweringonTruthfulQA
    % true (GPT-judge)· 2021-09-08
    27.17
    best: 53.24 (UnifiedQA 3B)
    TruthfulQA: Measuring How Models Mimic Human FalsehoodsarXiv:2109.07958
  • Question AnsweringonTruthfulQA
    BLEU· 2021-09-08
    -7.58
    best: -0.16 (UnifiedQA 3B)
    TruthfulQA: Measuring How Models Mimic Human FalsehoodsarXiv:2109.07958
  • Question AnsweringonTruthfulQA
    BLEURT· 2021-09-08
    -0.31
    best: 0.08 (UnifiedQA 3B)
    TruthfulQA: Measuring How Models Mimic Human FalsehoodsarXiv:2109.07958
  • Question AnsweringonTruthfulQA
    MC1· 2021-09-08
    0.2
    best: 0.59 (GPT-4 (RLHF))
    TruthfulQA: Measuring How Models Mimic Human FalsehoodsarXiv:2109.07958
  • Question AnsweringonTruthfulQA
    MC2· 2021-09-08
    0.36
    best: 0.75 (Mistral-7B-Instruct-v0.2 + TruthX)
    TruthfulQA: Measuring How Models Mimic Human FalsehoodsarXiv:2109.07958
  • Question AnsweringonTruthfulQA
    ROUGE· 2021-09-08
    -11.35
    best: 1.76 (UnifiedQA 3B)
    TruthfulQA: Measuring How Models Mimic Human FalsehoodsarXiv:2109.07958