© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    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. 13674
    Prev
    Next
    INDEX
    Explanations

    words related to the beginning or initiation of processes or actions

    oai_token-act-pair · gpt-3.5-turbo

    phrases that indicate the beginning of actions or events

    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
    gery
    -0.69
    mun
    -0.66
     anat
    -0.63
    ether
    -0.63
    que
    -0.63
    ocument
    -0.63
    iven
    -0.59
     Attribution
    -0.59
    owment
    -0.58
    ighed
    -0.57
    POSITIVE LOGITS
     starts
    3.34
     Starts
    2.63
     begins
    2.48
     start
    1.99
     finishes
    1.66
    start
    1.60
     Start
    1.59
     kicks
    1.58
     resumes
    1.58
     Begins
    1.57
    Activations Density 0.010%

    No Known Activations