© 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. 12341
    Prev
    Next
    INDEX
    Explanations

    words related to diseases, particularly malignant ones

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "mal" or its variations, often related to negative medical or ethical contexts

    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
    soDeliveryDate
    -0.92
    coming
    -0.77
     Eco
    -0.74
    ower
    -0.73
     Tablet
    -0.70
    ops
    -0.69
     Encounter
    -0.68
     Subaru
    -0.68
    skirts
    -0.66
    OPS
    -0.65
    POSITIVE LOGITS
     mal
    3.69
    mal
    1.87
    Mal
    1.63
     Mal
    1.63
     malt
    1.59
     mang
    1.18
     vulner
    1.14
     brut
    1.10
     tumor
    1.09
     Malt
    1.09
    Activations Density 0.023%

    No Known Activations