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. 8-RES-JB
    6. 5267
    Prev
    Next
    INDEX
    Explanations

    mentions of the television network HBO

    oai_token-act-pair · gpt-3.5-turbo

    mentions of the HBO network and its programming

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    uation
    -0.86
    acle
    -0.80
    acular
    -0.79
    alion
    -0.77
    ional
    -0.76
    forms
    -0.70
    ertodd
    -0.69
    jury
    -0.69
    uated
    -0.69
    ivated
    -0.68
    POSITIVE LOGITS
    EW
    0.88
     Wire
    0.86
    TV
    0.85
     Films
    0.84
     HBO
    0.84
    IGHTS
    0.81
     Networks
    0.78
    TW
    0.78
    MED
    0.77
    XM
    0.76
    Activations Density 0.022%

    No Known Activations