Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 0-RES-JB
    6. 3018
    Prev
    Next
    INDEX
    Explanations

    phrases related to boldness or emphasis

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "bold" and related concepts indicating strength or prominence

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    trap
    -0.67
     collects
    -0.65
    inary
    -0.65
    arium
    -0.62
    ector
    -0.62
    Newsletter
    -0.62
     Monitoring
    -0.61
     valve
    -0.59
    ania
    -0.59
    apor
    -0.59
    POSITIVE LOGITS
     bold
    3.62
    bold
    2.31
     Bold
    2.04
     ital
    1.84
     daring
    1.68
     boldly
    1.62
     brave
    1.33
     courageous
    1.32
     ambitious
    1.28
     brazen
    1.27
    Activations Density 0.021%

    No Known Activations