© 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 49152
    9. 8-RES_FS49152-JB
    11. 37301
    Prev
    Next
    INDEX
    Explanations

    structural elements or formatting indicators within the text

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    314
    +0.24
    1.1%
    640
    +0.17
    0.8%
    160
    +0.15
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    627
    +0.24
    0.04
    160
    +0.17
    0.04
    314
    +0.15
    0.01
    Negative Logits
     retard
    -0.81
     entitle
    -0.73
     tenants
    -0.72
     spatial
    -0.72
     creations
    -0.70
     ordinary
    -0.69
     disappear
    -0.68
     dens
    -0.66
     everyday
    -0.66
     furnish
    -0.64
    POSITIVE LOGITS
    pictured
    1.08
    formerly
    1.07
    tie
    1.05
    who
    1.03
    aka
    1.03
    pron
    1.01
    six
    1.01
    via
    1.00
    sports
    1.00
    four
    1.00
    Activations Density 0.052%

    No Known Activations