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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 1536
    5. 8-RES_FS1536-JB
    6. 579
    Prev
    Next
    INDEX
    Explanations

    dates and countries

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs1536-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    1,536
    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
    459
    +0.16
    0.7%
    284
    +0.11
    0.5%
    183
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.16
    0.18
    326
    +0.11
    0.16
    659
    +0.10
    0.09
    Negative Logits
    anooga
    -0.99
    plate
    -0.76
    essions
    -0.74
    pires
    -0.73
    raltar
    -0.72
    riages
    -0.71
    itals
    -0.70
    */(
    -0.70
    jriwal
    -0.69
    ieve
    -0.69
    POSITIVE LOGITS
    ors
    0.90
     Apostles
    0.82
     Vaugh
    0.74
    von
    0.71
    naire
    0.70
    enge
    0.69
    etta
    0.69
    ance
    0.67
     Flesh
    0.67
    ctr
    0.67
    Activations Density 3.586%

    No Known Activations