© 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. 20371
    Prev
    Next
    INDEX
    Explanations

    patterns that start with special characters and involve a sense or perception

    oai_token-act-pair · gpt-3.5-turbo

    special characters or glyphs that may indicate a certain tone or emphasis in the text

    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
     disse
    -0.79
     seiz
    -0.79
    ãĥ¯ãĥ³
    -0.76
     snail
    -0.71
     obser
    -0.71
     Franch
    -0.70
     scor
    -0.68
     hemor
    -0.65
     dehuman
    -0.64
    icit
    -0.63
    POSITIVE LOGITS
    Ŀ
    1.58
    ¡
    1.19
    Ĵ
    1.01
    ľ
    0.98
    ī
    0.98
    ¤
    0.97
    Ĩ
    0.95
    Ķ
    0.94
    ¦
    0.94
    ĺ
    0.92
    Activations Density 0.299%

    No Known Activations