© 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. 45727
    Prev
    Next
    INDEX
    Explanations

    terms that indicate quantity or reference to groups

    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
    291
    +0.10
    0.4%
    56
    +0.09
    0.4%
    102
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    163
    +0.10
    0.03
    313
    +0.09
    0.03
    216
    +0.09
    0.03
    Negative Logits
    etsk
    -0.76
    zyk
    -0.73
    arella
    -0.71
    verb
    -0.70
    agar
    -0.68
     chops
    -0.67
    way
    -0.66
    needs
    -0.66
    mere
    -0.65
    hran
    -0.65
    POSITIVE LOGITS
     successive
    0.85
     deceased
    0.82
     peoples
    0.80
    thood
    0.79
     individuals
    0.72
     Ancients
    0.72
     preceding
    0.69
    ankind
    0.68
     mankind
    0.66
     individual
    0.66
    Activations Density 0.145%

    No Known Activations