© 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. 13076
    Prev
    Next
    INDEX
    Explanations

    phrases indicating significant importance or impact

    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
    195
    +0.11
    0.5%
    386
    +0.10
    0.5%
    762
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    737
    +0.11
    0.02
    243
    +0.10
    0.01
    762
    +0.10
    0.01
    Negative Logits
    asus
    -0.87
    etheus
    -0.85
    elve
    -0.77
    perty
    -0.74
    ©¶æ
    -0.74
    oreAnd
    -0.72
    raltar
    -0.72
    icc
    -0.71
    avorite
    -0.71
    everal
    -0.71
    POSITIVE LOGITS
    breaker
    0.97
     breaker
    0.91
    killer
    0.90
    killers
    0.80
    deals
    0.78
    ership
    0.77
    ually
    0.76
    atical
    0.72
     deal
    0.69
     disadvant
    0.68
    Activations Density 0.013%

    No Known Activations