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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 1014
    Prev
    Next
    INDEX
    Explanations

    political and governmental terms and activities

    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
    373
    +0.11
    0.5%
    572
    +0.10
    0.5%
    630
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.11
    0.11
    451
    +0.10
    0.07
    603
    +0.10
    0.08
    Negative Logits
    raq
    -0.79
    heres
    -0.73
    lag
    -0.68
     Cruel
    -0.67
    imgur
    -0.65
     bugs
    -0.63
    rencies
    -0.63
    products
    -0.62
     caches
    -0.61
     Contains
    -0.61
    POSITIVE LOGITS
     helm
    1.19
     roles
    1.17
     role
    1.13
     presidency
    1.13
     duties
    1.07
     nomination
    1.03
     honorary
    0.98
     leadership
    0.98
    role
    0.96
     chair
    0.95
    Activations Density 3.516%

    No Known Activations