© 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. 2063
    Prev
    Next
    INDEX
    Explanations

    words related to the start or initiation of events

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "began."

    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
     veto
    -0.69
     exceptions
    -0.64
    oret
    -0.61
     persuasion
    -0.58
     behalf
    -0.56
    ether
    -0.55
     context
    -0.55
    vote
    -0.54
     specific
    -0.53
    ict
    -0.52
    POSITIVE LOGITS
     began
    3.21
     started
    2.36
     commenced
    2.31
     begun
    2.16
     begins
    2.15
    started
    1.87
     begin
    1.85
     resumed
    1.80
     ceased
    1.80
     sprang
    1.72
    Activations Density 0.021%

    No Known Activations