© 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. 13319
    Prev
    Next
    INDEX
    Explanations

    references to art

    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
    14
    +0.11
    0.5%
    288
    +0.10
    0.4%
    670
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    670
    +0.11
    0.02
    414
    +0.10
    0.02
    523
    +0.10
    0.02
    Negative Logits
     htt
    -0.77
     Wyoming
    -0.69
     XT
    -0.68
     DN
    -0.66
     Crosby
    -0.64
     Ingram
    -0.63
     sshd
    -0.63
    KI
    -0.62
     Tire
    -0.61
     Witnesses
    -0.61
    POSITIVE LOGITS
    istry
    1.64
    isans
    1.58
    ifice
    1.43
    emis
    1.41
    esian
    1.41
    ifacts
    1.30
    works
    1.30
    illery
    1.29
    isan
    1.19
    ificial
    1.15
    Activations Density 0.025%

    No Known Activations