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. 113
    Prev
    Next
    INDEX
    Explanations

    sports-related terms and concepts

    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
    288
    +0.12
    0.5%
    490
    +0.10
    0.5%
    324
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.12
    0.25
    575
    +0.10
    0.18
    324
    +0.09
    0.11
    Negative Logits
     channels
    -0.89
     Machines
    -0.87
    eworks
    -0.86
    tics
    -0.86
    sites
    -0.84
     establishments
    -0.83
     sights
    -0.82
    ernels
    -0.82
    rooms
    -0.81
    casts
    -0.81
    POSITIVE LOGITS
     standpoint
    0.98
     consisting
    0.89
     resembling
    0.87
     Called
    0.86
     akin
    0.84
     called
    0.82
     reminiscent
    0.81
     spokeswoman
    0.80
    atical
    0.79
     spokesperson
    0.78
    Activations Density 4.342%

    No Known Activations