© 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. 45763
    Prev
    Next
    INDEX
    Explanations

    words related to spatial positioning and movement

    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
    496
    +0.14
    0.6%
    556
    +0.11
    0.5%
    762
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    556
    +0.14
    0.01
    496
    +0.11
    0.02
    585
    +0.10
    0.01
    Negative Logits
    ancial
    -0.79
    Bal
    -0.73
    antine
    -0.71
    unin
    -0.69
    cients
    -0.68
    accompan
    -0.68
    allo
    -0.67
    Fourth
    -0.67
    anos
    -0.66
    emporary
    -0.65
    POSITIVE LOGITS
    xual
    0.71
    EMENT
    0.71
    println
    0.69
    dden
    0.68
    ned
    0.67
    stretched
    0.63
     loud
    0.60
     puff
    0.59
    oeuv
    0.58
    =\"
    0.58
    Activations Density 0.043%

    No Known Activations