© 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. 4-RES-JB
    11. 17309
    Prev
    Next
    INDEX
    Explanations

    specific names, likely proper nouns

    oai_token-act-pair · gpt-3.5-turbo

    names of individuals and characters associated with significant events or contexts

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

    No Comments

    Negative Logits
    journal
    -0.85
     nerv
    -0.69
     saline
    -0.68
    arenthood
    -0.68
     governors
    -0.67
    igators
    -0.65
    izoph
    -0.65
    oxy
    -0.63
     captcha
    -0.63
    ggy
    -0.62
    POSITIVE LOGITS
     Curtis
    0.80
     Compton
    0.78
    hift
    0.74
     Yar
    0.74
     Dixon
    0.72
     Armstrong
    0.69
     McKenzie
    0.68
    aba
    0.68
    ensor
    0.67
    ãĤ¼ãĤ¦ãĤ¹
    0.67
    Activations Density 0.010%

    No Known Activations