© 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. 26546
    Prev
    Next
    INDEX
    Explanations

    references to the 21st century and significant events or trends associated with it

    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
    +0.14
    0.6%
    223
    +0.12
    0.5%
    151
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    223
    +0.14
    0.04
    693
    +0.12
    0.04
    310
    +0.11
    0.04
    Negative Logits
     externalToEVAOnly
    -0.83
    aneous
    -0.75
    ensional
    -0.73
    ulously
    -0.71
     Polo
    -0.69
    backs
    -0.68
    riages
    -0.66
    atics
    -0.66
    tered
    -0.64
     gren
    -0.64
    POSITIVE LOGITS
    st
    1.57
    ST
    0.84
    nect
    0.79
    worthy
    0.75
    ĨĴ
    0.73
    onsense
    0.72
    âģ
    0.70
    Counter
    0.70
    istries
    0.70
    entimes
    0.69
    Activations Density 0.056%

    No Known Activations