© 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. 4407
    Prev
    Next
    INDEX
    Explanations

    phrases asking for trust or subscriptions

    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
    551
    +0.12
    0.5%
    49
    +0.11
    0.5%
    107
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    551
    +0.12
    0.05
    49
    +0.11
    0.04
    107
    +0.10
    0.03
    Negative Logits
     Politics
    -0.69
     Cheong
    -0.67
     deteriorated
    -0.65
     Mant
    -0.65
     Sons
    -0.63
     Americas
    -0.63
     Deer
    -0.61
     Tsuk
    -0.61
     Unch
    -0.61
    phas
    -0.61
    POSITIVE LOGITS
    't
    1.12
    NOT
    1.03
     emulate
    1.02
     afford
    0.91
     relate
    0.88
     use
    0.88
     utilize
    0.83
     distribute
    0.81
    ¶ħ
    0.80
    intend
    0.80
    Activations Density 0.074%

    No Known Activations