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 1536
    5. 8-RES_FS1536-JB
    6. 1428
    Prev
    Next
    INDEX
    Explanations

    names of sports teams and players

    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
    602
    +0.11
    0.5%
    68
    +0.11
    0.5%
    611
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.11
    0.23
    752
    +0.11
    0.17
    67
    +0.11
    0.15
    Negative Logits
    agame
    -0.83
    iphate
    -0.79
    anse
    -0.76
    ructure
    -0.75
    swer
    -0.75
    ouf
    -0.73
    undrum
    -0.73
    epad
    -0.72
    bsite
    -0.72
    ixir
    -0.72
    POSITIVE LOGITS
     officials
    1.14
    goers
    1.05
     audiences
    0.98
     policymakers
    0.97
     administrators
    0.93
     opponents
    0.93
     staffers
    0.92
     emergencies
    0.91
     sources
    0.91
     insiders
    0.91
    Activations Density 3.993%

    No Known Activations