© 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. 3-RES-JB
    11. 13148
    Prev
    Next
    INDEX
    Explanations

    words related to capturing or seizing something or someone

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "capture" and its variations

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     notor
    -0.77
     amen
    -0.76
     conflic
    -0.71
     occas
    -0.70
    senal
    -0.65
     inclined
    -0.65
    gom
    -0.63
     err
    -0.63
     millenn
    -0.62
    AE
    -0.62
    POSITIVE LOGITS
    racuse
    0.91
    nect
    0.79
     captured
    0.77
    otes
    0.73
    doors
    0.72
    ption
    0.71
     recapt
    0.70
    oufl
    0.70
     capture
    0.70
    oons
    0.69
    Activations Density 0.016%

    No Known Activations