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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 4-RES-JB
    11. 19549
    Prev
    Next
    INDEX
    Explanations

    mentions of names, particularly with the name "Dre" being the most prominent

    oai_token-act-pair · gpt-3.5-turbo

    proper nouns, particularly names of people

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

    No Comments

    Negative Logits
     Controlled
    -0.67
     Uganda
    -0.65
     Sark
    -0.64
    guide
    -0.61
    UI
    -0.61
     Ulster
    -0.57
     Penal
    -0.57
     Administrative
    -0.57
     Tunisia
    -0.56
    leaders
    -0.56
    POSITIVE LOGITS
     Dre
    0.92
    fters
    0.92
    iband
    0.83
    essen
    0.83
    anguage
    0.78
     fumes
    0.75
    ugh
    0.74
    éĹĺ
    0.74
    wcs
    0.72
    mson
    0.72
    Activations Density 0.020%

    No Known Activations