© 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. 14592
    Prev
    Next
    INDEX
    Explanations

    references to demonstrating or demonstration

    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
    167
    +0.11
    0.5%
    288
    +0.11
    0.5%
    665
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    724
    +0.11
    0.03
    75
    +0.11
    0.02
    331
    +0.10
    0.02
    Negative Logits
    76561
    -0.85
    lake
    -0.77
    agues
    -0.74
    cloth
    -0.73
    ita
    -0.69
    osate
    -0.69
    oldown
    -0.68
    pool
    -0.68
    turn
    -0.68
    tta
    -0.68
    POSITIVE LOGITS
     feats
    0.90
    antly
    0.81
     prowess
    0.81
     displeasure
    0.78
     excellence
    0.77
     proficiency
    0.73
    anooga
    0.72
    iveness
    0.70
     how
    0.70
    ivity
    0.70
    Activations Density 0.040%

    No Known Activations