© 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. 2324
    Prev
    Next
    INDEX
    Explanations

    strings of numbers that seem like codes or identifiers

    oai_token-act-pair · gpt-3.5-turbo

    numerical values, particularly dates and counts

    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
    •Embedding Space Digit Features

    No Comments

    Negative Logits
     Spock
    -0.68
    elo
    -0.68
    uffle
    -0.66
     Celestial
    -0.64
     Ler
    -0.63
     Squ
    -0.62
     burner
    -0.61
     Leonardo
    -0.61
    ello
    -0.61
    socket
    -0.60
    POSITIVE LOGITS
    21
    1.90
    41
    1.89
    81
    1.78
    61
    1.77
    51
    1.71
    71
    1.60
    171
    1.53
    141
    1.49
    131
    1.47
    91
    1.47
    Activations Density 0.057%

    No Known Activations