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. 43177
    Prev
    Next
    INDEX
    Explanations

    names of people

    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
    610
    +0.12
    0.5%
    66
    +0.10
    0.5%
    708
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.12
    0.07
    700
    +0.10
    0.06
    458
    +0.09
    0.05
    Negative Logits
    PDATE
    -0.81
    GoldMagikarp
    -0.77
     rall
    -0.73
     looph
    -0.71
    idated
    -0.67
    ÃĥÃĤ
    -0.65
    uesday
    -0.65
     erection
    -0.63
    translation
    -0.63
    Motor
    -0.63
    POSITIVE LOGITS
     herself
    1.59
    eva
    1.05
    ova
    1.02
     husband
    0.98
     Actress
    0.94
     breasts
    0.93
     Christina
    0.93
    kaya
    0.90
     hijab
    0.88
     She
    0.87
    Activations Density 0.204%

    No Known Activations