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. 8435
    Prev
    Next
    INDEX
    Explanations

    Japanese names or terms

    oai_token-act-pair · gpt-3.5-turboTriggered 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
    480
    +0.29
    1.4%
    481
    +0.18
    0.9%
    447
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.29
    0.06
    579
    +0.18
    0.06
    401
    +0.10
    0.06
    Negative Logits
    romeda
    -0.62
    .''.
    -0.61
    .�
    -0.58
    .):
    -0.57
    %);
    -0.56
     revolving
    -0.56
    )",
    -0.54
    ''.
    -0.54
    ";
    -0.54
     ����
    -0.53
    POSITIVE LOGITS
     has
    0.86
     intends
    0.81
     Productions
    0.81
     employs
    0.80
     succeeds
    0.78
     Ltd
    0.78
     understands
    0.77
     operates
    0.77
     considers
    0.77
     represents
    0.76
    Activations Density 0.530%

    No Known Activations