© 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. 756
    Prev
    Next
    INDEX
    Explanations

    references to tall structures or buildings

    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
    288
    +0.10
    0.5%
    127
    +0.10
    0.5%
    565
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    565
    +0.10
    0.02
    99
    +0.10
    0.01
    288
    +0.10
    0.02
    Negative Logits
    lde
    -0.76
    ktop
    -0.75
    eer
    -0.75
    actionDate
    -0.73
    vernment
    -0.72
    arty
    -0.72
     Wonderland
    -0.70
    BOOK
    -0.70
    в
    -0.68
    eering
    -0.68
    POSITIVE LOGITS
     stature
    1.13
     tallest
    0.98
     taller
    0.92
     towers
    0.90
    falls
    0.86
    ened
    0.85
    itudinal
    0.82
     ceilings
    0.81
    mad
    0.81
    itect
    0.80
    Activations Density 0.062%

    No Known Activations