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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 45812
    Prev
    Next
    INDEX
    Explanations

    words associated with risks, consequences, and the importance of safety in various contexts

    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
    480
    +0.42
    2.1%
    447
    +0.15
    0.8%
    266
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.42
    0.07
    455
    +0.15
    0.06
    480
    +0.11
    0.01
    Negative Logits
    selves
    -0.64
    ovember
    -0.60
    enegger
    -0.59
    olulu
    -0.57
    +.
    -0.56
    poon
    -0.55
    iolet
    -0.54
    ornings
    -0.54
     Ago
    -0.54
    ECA
    -0.54
    POSITIVE LOGITS
    iest
    0.81
     varies
    0.80
     becomes
    0.75
     consists
    0.72
     remains
    0.71
     is
    0.71
     goes
    0.69
     itself
    0.69
     reaches
    0.67
     disappears
    0.67
    Activations Density 0.276%

    No Known Activations