Capability gains

GPT-3 demonstrates broad few-shot learning at 175 billion parameters

OpenAI's original GPT-3 paper reported a 175-billion-parameter autoregressive language model that performed many tasks from instructions or a few examples without gradient updates, while also documenting weaknesses, bias, and human difficulty distinguishing some generated news. OpenAI separately deployed GPT-3-family weights through a controlled private-beta API on June 11.

0 comments · 0 votesOpen discussion

Public discussion is readable by everyone. Sign in to comment, reply, or vote.

No comments yet. Start the discussion.

CURRENT ASSESSMENT · REVISION 1
TOWARD DOOM47confidence 90/100

Why it moved the index

GPT-3 established a major general-purpose capability and scaling milestone, including broad task adaptation and realistic text generation, and its separately dated API deployment verified practical downstream use. Controlled access and low independent agency moderate the assessment.

AUDIT TRAIL

Assessment history

  1. R1
    Toward 47 · confidence 90

    New original research milestone with separately verified practical deployment, exact identity, and no durable duplicate.

    11 Aug 2026
SHARE THE FINDINGS

Share this page

DoomBench social sharing card for GPT-3 demonstrates broad few-shot learning at 175 billion parameters.
  1. DoomBench assesses “GPT-3 demonstrates broad few-shot learning at 175 billion parameters” as evidence moving toward doom, with magnitude 47 and confidence 90 out of 100 in the capability gains category.

  2. The DoomBench assessment of “GPT-3 demonstrates broad few-shot learning at 175 billion parameters” is based on reporting from arXiv and records the editorial rationale, source quality, attribution, and revision history.

  3. DoomBench summarizes “GPT-3 demonstrates broad few-shot learning at 175 billion parameters” as follows: OpenAI's original GPT-3 paper reported a 175-billion-parameter autoregressive language model that performed many tasks from...

    https://www.doombench.com/news/gpt-3-demonstrates-broad-few-shot-learning-at-175-billion-parameters-2020-05-28