© 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. 1-RES-JB
    11. 16093
    Prev
    Next
    INDEX
    Explanations

    references to titles or individuals with "Lords" or similar noble titles

    oai_token-act-pair · gpt-3.5-turbo

    references to titles of nobility or authority

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    icago
    -0.76
    OPLE
    -0.70
     SAP
    -0.70
     Spokane
    -0.65
     fir
    -0.65
    Gaza
    -0.65
    PsyNetMessage
    -0.65
     Bie
    -0.64
    ocaust
    -0.64
     Helsinki
    -0.63
    POSITIVE LOGITS
     Lords
    1.29
    hip
    0.99
    hips
    0.94
     lords
    0.93
    lords
    0.91
     Mods
    0.79
    ync
    0.79
    rats
    0.77
    masters
    0.76
     Halls
    0.75
    Activations Density 0.017%

    No Known Activations