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 24576
    5. 8-RES_FS24576-JB
    6. 23112
    Prev
    Next
    INDEX
    Explanations

    references to Algeria and Tunisia

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    481
    +0.15
    0.7%
    288
    +0.14
    0.6%
    447
    +0.14
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.15
    0.02
    612
    +0.14
    0.02
    67
    +0.14
    0.02
    Negative Logits
    bilt
    -0.86
    ramid
    -0.77
    rint
    -0.75
    pling
    -0.72
    verett
    -0.70
    BY
    -0.70
    onz
    -0.70
    rote
    -0.69
    ertodd
    -0.69
    ersive
    -0.69
    POSITIVE LOGITS
    ian
    1.08
    ians
    0.88
     Tunisia
    0.88
    arian
    0.84
    ienne
    0.82
    ique
    0.80
     Tunis
    0.79
     Mali
    0.78
     Tanz
    0.77
    ican
    0.76
    Activations Density 0.022%

    No Known Activations