© 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 12288
    9. 8-RES_FS12288-JB
    11. 351
    Prev
    Next
    INDEX
    Explanations

    phrases related to events or protests

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    57
    +0.16
    0.8%
    167
    +0.11
    0.5%
    85
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    57
    +0.16
    0.03
    85
    +0.11
    0.03
    232
    +0.11
    0.03
    Negative Logits
     Grande
    -0.69
     mosaic
    -0.65
     Awakening
    -0.65
    phrine
    -0.64
     Letter
    -0.64
    ITAL
    -0.63
    ULAR
    -0.63
    WARE
    -0.63
    uated
    -0.62
    ulously
    -0.61
    POSITIVE LOGITS
    istle
    1.33
    irlwind
    1.22
    olly
    1.22
    ilst
    1.20
    irling
    1.20
    olen
    1.17
    omever
    1.15
    irl
    1.13
    izzard
    1.13
    izz
    1.11
    Activations Density 0.054%

    No Known Activations