© 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. 3-RES-JB
    11. 4713
    Prev
    Next
    INDEX
    Explanations

    words related to genitalia, specifically focusing on the term "penis"

    oai_token-act-pair · gpt-3.5-turbo

    references to male and female genitalia

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    quickShipAvailable
    -0.90
    PsyNetMessage
    -0.76
    arily
    -0.75
    HCR
    -0.70
    ģĸ
    -0.69
    WER
    -0.68
    Trend
    -0.68
    ONSORED
    -0.65
    occ
    -0.64
     Volunte
    -0.63
    POSITIVE LOGITS
     penis
    1.09
     ejac
    0.98
     enlarg
    0.93
     dick
    0.89
     anus
    0.85
     shaft
    0.85
     butt
    0.82
    hole
    0.80
     canal
    0.80
    holes
    0.78
    Activations Density 0.007%

    No Known Activations