© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 0-RES-JB
    11. 13718
    Prev
    Next
    INDEX
    Explanations

    mentions of supernatural practices or dark arts, specifically witchcraft or sorcery

    oai_token-act-pair · gpt-3.5-turbo

    terms associated with magical or supernatural practices

    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
    emort
    -0.76
    agi
    -0.75
     emperor
    -0.69
    icide
    -0.67
    untu
    -0.66
    theless
    -0.65
     acid
    -0.65
    ":["
    -0.64
    acters
    -0.64
    iyah
    -0.63
    POSITIVE LOGITS
    ORN
    0.75
     Ways
    0.68
    WT
    0.68
    pter
    0.67
    METHOD
    0.66
    esse
    0.66
    FR
    0.65
     Bucket
    0.61
    liam
    0.60
    WOOD
    0.60
    Activations Density 0.000%

    No Known Activations