© 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 EvalsExportsAPI 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. 12174
    Prev
    Next
    INDEX
    Explanations

    phrases related to instructions or suggestions

    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
    124
    +0.10
    0.5%
    394
    +0.10
    0.5%
    278
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    38
    +0.10
    0.04
    195
    +0.10
    0.02
    70
    +0.09
    0.04
    Negative Logits
    ĸļ
    -0.77
     supposedly
    -0.61
    ulner
    -0.61
    unda
    -0.60
     ALWAYS
    -0.59
    Apps
    -0.58
    Tube
    -0.56
    jab
    -0.56
    ruction
    -0.56
     evidently
    -0.55
    POSITIVE LOGITS
     someday
    1.08
     depending
    0.78
     tempted
    0.77
    ivably
    0.72
     slightly
    0.72
     inadvertently
    0.71
    xus
    0.71
     underest
    0.70
     momentarily
    0.67
    depending
    0.67
    Activations Density 0.313%

    No Known Activations