© 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 · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 13546
    Prev
    Next
    INDEX
    Explanations

    specific references to a place or entity named "Thunder"

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    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

    Neuron Alignment
    Index
    Value
    % of L₁
    472
    +0.13
    0.6%
    167
    +0.11
    0.5%
    288
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.13
    0.01
    288
    +0.11
    0.01
    464
    +0.10
    0.01
    Negative Logits
    VICE
    -0.84
    LECT
    -0.75
    Constructed
    -0.66
    utions
    -0.63
    lections
    -0.63
    utable
    -0.63
    76561
    -0.62
    PsyNetMessage
    -0.62
    verage
    -0.62
    ationally
    -0.60
    POSITIVE LOGITS
    storms
    1.47
    storm
    1.26
    bolt
    1.24
    bol
    1.23
    stru
    1.22
    bird
    1.20
    birds
    1.14
    cloud
    1.01
    fuck
    1.00
    hawk
    0.95
    Activations Density 0.022%

    No Known Activations