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

    Neuronpedia

    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. 8-RES-JB
    11. 7843
    Prev
    Next
    INDEX
    Explanations

    instances related to sexual misconduct and assault

    oai_token-act-pair · gpt-3.5-turbo

    references to sexual assault and misconduct allegations

    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
    pmwiki
    -0.86
    NetMessage
    -0.82
     certs
    -0.76
     cancell
    -0.71
    inventoryQuantity
    -0.68
     pronounce
    -0.67
    annis
    -0.67
    govtrack
    -0.66
    ciation
    -0.66
    ãĥķãĤ©
    -0.65
    POSITIVE LOGITS
     underage
    1.33
     sexually
    1.05
     inappropriately
    1.02
     genitals
    1.01
     prostitute
    0.99
     prostitutes
    0.99
     sexual
    0.97
    girlfriend
    0.96
     minors
    0.96
    Sexual
    0.93
    Activations Density 0.348%

    No Known Activations