© 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. 7597
    Prev
    Next
    INDEX
    Explanations

    racial terms, specifically focusing on identifying mentions of Black and White individuals in the text

    oai_token-act-pair · gpt-3.5-turbo

    references to racial groups and discussions about race relations

    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
     template
    -0.65
    ument
    -0.65
    acher
    -0.65
    EG
    -0.64
     sandbox
    -0.63
    oven
    -0.62
     signature
    -0.62
    wallet
    -0.62
    ö
    -0.62
     standalone
    -0.61
    POSITIVE LOGITS
     Blacks
    4.16
     blacks
    2.45
     Whites
    2.13
     Negro
    1.68
     Africans
    1.60
     negro
    1.52
     Asians
    1.50
     Hispanics
    1.49
     Black
    1.32
     Caucas
    1.26
    Activations Density 0.015%

    No Known Activations