© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 17322
    Prev
    Next
    INDEX
    Explanations

    the adjective "pretty" with a high activation value

    oai_token-act-pair · gpt-3.5-turbo

    the word "pretty" and its variations, indicating a focus on subjective evaluations or modifiers of intensity in descriptions

    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
    venant
    -0.85
    iment
    -0.80
    arta
    -0.80
    PLIED
    -0.78
    iments
    -0.75
    aer
    -0.72
    herry
    -0.72
    APD
    -0.70
    uther
    -0.69
    pent
    -0.68
    POSITIVE LOGITS
     darn
    1.04
     nifty
    0.81
     hairy
    0.76
     damn
    0.75
     harmless
    0.74
     cool
    0.70
     innocuous
    0.70
     noses
    0.69
     neat
    0.69
     snipp
    0.68
    Activations Density 0.018%

    No Known Activations