© 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 EvalsExportsinterp-engineNEWAPI 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. 8-RES-JB
    11. 12398
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "pro" followed by another portion of a word

    oai_token-act-pair · gpt-3.5-turbo

    references to pro-choice themes and advocacy

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ãĤ¼ãĤ¦ãĤ¹
    -0.83
     needles
    -0.74
     spears
    -0.69
     straw
    -0.69
     Sins
    -0.67
    assetsadobe
    -0.67
    ĸļ
    -0.66
     Halls
    -0.66
     strawberries
    -0.65
     halls
    -0.65
    POSITIVE LOGITS
    blems
    1.52
    ceed
    1.39
    gression
    1.32
    digy
    1.28
    secut
    1.27
    gressive
    1.27
    spective
    1.26
    secution
    1.22
    spect
    1.21
    posal
    1.18
    Activations Density 0.024%

    No Known Activations