© 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. 0-RES-JB
    11. 15447
    Prev
    Next
    INDEX
    Explanations

    mentions of portrait photography or artistic representations of individuals

    oai_token-act-pair · gpt-3.5-turbo

    elements related to specific traits and characteristics of individuals, particularly focusing on the term "rait" and names associated with Hardy

    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
    DOS
    -0.75
     cruising
    -0.69
     artific
    -0.66
     synt
    -0.63
     Morg
    -0.62
     Vern
    -0.62
    teenth
    -0.62
    ften
    -0.61
     curs
    -0.60
     DOS
    -0.60
    POSITIVE LOGITS
    rait
    2.35
    raits
    1.76
     Abbey
    1.75
    rito
    1.67
     Hardy
    1.50
    itutional
    1.19
    ippi
    1.11
    mallow
    1.08
    nery
    1.07
    mary
    1.03
    Activations Density 0.052%

    No Known Activations