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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 768
    5. 8-RES_FS768-JB
    6. 356
    Prev
    Next
    INDEX
    Explanations

    names and titles of individuals

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    590
    +0.11
    0.5%
    431
    +0.10
    0.4%
    709
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    57
    +0.11
    0.08
    590
    +0.10
    0.08
    659
    +0.09
    0.09
    Negative Logits
    theless
    -0.73
    schild
    -0.66
     favour
    -0.61
     rud
    -0.60
     evenings
    -0.59
     caution
    -0.57
     discour
    -0.55
     Mish
    -0.54
     whisk
    -0.54
    adobe
    -0.54
    POSITIVE LOGITS
    .,
    1.94
    .?
    1.66
    .;
    1.48
    .:
    1.47
    .,"
    1.46
    ./
    1.30
    .),
    1.19
    .—
    1.16
    orporated
    1.15
    .-
    1.12
    Activations Density 2.756%

    No Known Activations