© 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. 2415
    Prev
    Next
    INDEX
    Explanations

    words relating to majority or prevalence in a specific demographic or geographic context

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating dominance or majority within a demographic context

    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
    dimensional
    -0.77
    senal
    -0.66
     confinement
    -0.65
     millenn
    -0.63
    conference
    -0.62
    aditional
    -0.61
     Settlement
    -0.61
    stellar
    -0.61
     Granger
    -0.60
    Graph
    -0.60
    POSITIVE LOGITS
     predomin
    1.79
     predominantly
    0.81
    âĨij
    0.69
    ér
    0.69
    istries
    0.67
    ieu
    0.67
    neys
    0.66
    ces
    0.62
    ues
    0.62
    culation
    0.61
    Activations Density 0.004%

    No Known Activations