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. 1900
    Prev
    Next
    INDEX
    Explanations

    negative sentiments and criticisms towards specific legislation

    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
    605
    +0.11
    0.5%
    586
    +0.10
    0.5%
    652
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    575
    +0.11
    0.12
    564
    +0.10
    0.07
    259
    +0.10
    0.05
    Negative Logits
     Cosponsors
    -0.80
    Ĥİ
    -0.80
    Benz
    -0.70
    ctors
    -0.70
    advertising
    -0.69
    WARD
    -0.65
    ãĤ¹ãĥĪ
    -0.63
     ACTIONS
    -0.62
    inka
    -0.62
    ugh
    -0.62
    POSITIVE LOGITS
     lined
    1.02
     installed
    0.86
     ready
    0.86
     backing
    0.83
     tow
    0.82
     attached
    0.82
     handy
    0.81
     tucked
    0.80
     strapped
    0.79
    installed
    0.78
    Activations Density 2.282%

    No Known Activations