RLTasksby ulam.ai
Models / OpenAI

OpenAI / MODEL EVIDENCE

GPT-2 (1.5B)

7 benchmarks · 10 observations · 4 sources · released 2019-11-05

Scores at a glance

7 benchmarks · large dot: GPT-2 (1.5B); small dots: closest models where reported

ARC AI2 ↗25%
23.244.9

GPT-2 (1.5B) · ARC AI2 · Challenge score · % ↗

PIQA ↗70.5%
6979.4

GPT-2 (1.5B) · PIQA · Score · % ↗

Closest models by score

Similar positions on shared benchmark leaderboards. Smaller gaps are closer.

13.7 percentile-point gap

4 of 6 shared benchmark comparisons

13.7 percentile-point gap

4 of 6 shared benchmark comparisons

How similarity is measured

Mean absolute gap in percentile positions on shared benchmark settings; lower is closer. At least 60% overlap is required, with a minimum of three when available. Independent evaluations are used when there are at least three; otherwise published report comparisons are used. Repeated model observations within one setting are averaged before ranking. Writing-style traits and composite indices are excluded. This describes similarity, not an overall quality ranking.

Filter comparisons to this model →

Sources

Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.

  1. Textbooks Are All You Need II: phi-1.5 technical report ↗

    Li, Yuanzhi, Bubeck, Sébastien, Eldan, Ronen, Del Giorno, Allie, Gunasekar, Suriya, Lee, Yin Tat · 2023

  2. WinoGrande ↗

    Epoch AI

  3. TruthfulQA: Measuring How Models Mimic Human Falsehoods — published result table ↗

    Lin, Stephanie, Hilton, Jacob, Evans, Owain · 2021

  4. Measuring AI Ability to Complete Long Software Tasks ↗

    METR

Names used in source tables

GPT-2 (1.5B) · GPT-2 1.5B