Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 26075
    Prev
    Next
    INDEX
    Explanations

    various forms of the root word "ra" or similar phonetic sounds

    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
    373
    +0.20
    0.9%
    745
    +0.16
    0.7%
    466
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    373
    +0.20
    -0.00
    466
    +0.16
    0.02
    232
    +0.10
    0.02
    Negative Logits
     Yemeni
    -0.65
     Samar
    -0.64
    ROR
    -0.63
    Demand
    -0.61
    QUI
    -0.60
     Dragonbound
    -0.59
     Duterte
    -0.59
     Reincarn
    -0.58
    ECTION
    -0.58
    PRESS
    -0.58
    POSITIVE LOGITS
    itzer
    1.06
    auer
    0.90
    itz
    0.90
    zinski
    0.89
    inski
    0.85
    enei
    0.84
    enburg
    0.83
    ussia
    0.83
    cko
    0.82
    enstein
    0.80
    Activations Density 0.020%

    No Known Activations