© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 7203
    Prev
    Next
    INDEX
    Explanations

    instances of the word "before" followed by either a number or a pronoun

    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
    59
    +0.10
    0.5%
    409
    +0.10
    0.4%
    90
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    59
    +0.10
    0.04
    90
    +0.10
    0.03
    88
    +0.09
    0.03
    Negative Logits
    aez
    -0.75
    Force
    -0.74
    SA
    -0.71
    Pool
    -0.70
    atic
    -0.70
    Eye
    -0.66
    ANA
    -0.66
    DC
    -0.65
    ãĤ¤
    -0.64
    atana
    -0.63
    POSITIVE LOGITS
     WWII
    0.81
     realizing
    0.74
     puberty
    0.73
     existed
    0.72
     anyone
    0.71
     dawn
    0.69
    vernment
    0.68
     sunrise
    0.66
    uberty
    0.66
    hower
    0.65
    Activations Density 0.065%

    No Known Activations