© 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. 26585
    Prev
    Next
    INDEX
    Explanations

    references to names and titles

    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
    377
    +0.09
    0.4%
    264
    +0.09
    0.4%
    88
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    264
    +0.09
    0.02
    131
    +0.09
    0.02
    689
    +0.09
    0.02
    Negative Logits
    igsaw
    -0.85
    onential
    -0.75
    ocent
    -0.69
    ython
    -0.68
    artifacts
    -0.67
    restling
    -0.67
    aza
    -0.66
    anga
    -0.66
    upiter
    -0.66
     Flavoring
    -0.65
    POSITIVE LOGITS
    )[
    0.81
     hereafter
    0.72
     moniker
    0.72
    displayText
    0.69
    ÑĮ
    0.68
    iHUD
    0.66
     wont
    0.65
    IAS
    0.65
    phrase
    0.64
    pron
    0.64
    Activations Density 0.056%

    No Known Activations