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 3072
    5. 8-RES_FS3072-JB
    6. 1037
    Prev
    Next
    INDEX
    Explanations

    phrases related to migration and social policies

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    720
    +0.12
    0.5%
    686
    +0.10
    0.5%
    431
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.12
    0.14
    686
    +0.10
    0.09
    720
    +0.09
    0.11
    Negative Logits
     Tarant
    -0.70
    framework
    -0.69
    player
    -0.67
    ãĥīãĥ©ãĤ´ãĥ³
    -0.66
    catentry
    -0.65
    REDACTED
    -0.64
    timer
    -0.63
    ãĥ´ãĤ¡
    -0.63
    DK
    -0.63
    sama
    -0.62
    POSITIVE LOGITS
     themselves
    1.27
     their
    1.18
    their
    1.12
    selves
    1.02
     THEIR
    1.00
     lifestyles
    0.95
     careers
    0.95
     caregivers
    0.92
     subsistence
    0.90
     disabilities
    0.87
    Activations Density 3.628%

    No Known Activations