© 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. 0-RES-JB
    11. 1354
    Prev
    Next
    INDEX
    Explanations

    special characters in a specific sequence

    oai_token-act-pair · gpt-3.5-turbo

    symbols and formatting elements within the text

    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
     Georgetown
    -0.84
     WikiLeaks
    -0.77
    geries
    -0.74
    gery
    -0.73
     Hercules
    -0.71
     CIA
    -0.69
     Liberty
    -0.69
     Barton
    -0.68
     Revenge
    -0.68
     Fargo
    -0.68
    POSITIVE LOGITS
    ã
    3.02
    âĸ
    2.57
    âĹ
    2.47
    âĹı
    2.42
    âĻ
    2.30
    âĶ
    2.25
    âĸº
    2.22
    Û
    2.17
    âĺ
    2.16
    âĪ
    2.05
    Activations Density 0.015%

    No Known Activations