© 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. 37545
    Prev
    Next
    INDEX
    Explanations

    occurrences of the character "Ãĸ" and variations of 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
    48
    +0.11
    0.5%
    420
    +0.10
    0.5%
    699
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    48
    +0.11
    0.01
    699
    +0.10
    0.01
    659
    +0.10
    0.01
    Negative Logits
    raints
    -0.78
     Keeper
    -0.76
     guiActiveUnfocused
    -0.73
    iflower
    -0.72
    aneously
    -0.72
    itud
    -0.72
     Sussex
    -0.70
     BDS
    -0.67
     totality
    -0.67
    itude
    -0.66
    POSITIVE LOGITS
    zbek
    1.02
    hl
    1.01
    ÄŁ
    0.97
    nder
    0.97
    conom
    0.91
    thing
    0.91
    sten
    0.91
    meric
    0.90
    hler
    0.89
    ternity
    0.87
    Activations Density 0.004%

    No Known Activations