© 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. 3-RES-JB
    11. 15944
    Prev
    Next
    INDEX
    Explanations

    special characters or codes, potentially indicating a specific type of formatting or language

    oai_token-act-pair · gpt-3.5-turbo

    special characters and symbols, particularly the cent sign

    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
    ulative
    -0.81
     Gutenberg
    -0.78
     Spiel
    -0.77
    gdala
    -0.76
    emonium
    -0.75
    enegger
    -0.74
    ciation
    -0.74
     Occupations
    -0.69
    coni
    -0.69
     Addiction
    -0.68
    POSITIVE LOGITS
    ¢
    0.98
    ĥ
    0.95
    ¨
    0.91
    ׾
    0.86
    ĭ
    0.84
    âĶĢâĶĢ
    0.84
    ï¸ı
    0.82
    ¦
    0.81
    §
    0.81
    «
    0.80
    Activations Density 0.005%

    No Known Activations