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. 2643
    Prev
    Next
    INDEX
    Explanations

    phrases expressing opinions or viewpoints

    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
    218
    +0.12
    0.5%
    497
    +0.11
    0.5%
    541
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    541
    +0.12
    0.08
    102
    +0.11
    0.09
    647
    +0.11
    0.07
    Negative Logits
     answ
    -0.71
    mentioned
    -0.68
    ramids
    -0.67
    then
    -0.64
    aunted
    -0.63
    cies
    -0.63
    ersen
    -0.62
    rote
    -0.62
    ITH
    -0.62
    leaf
    -0.61
    POSITIVE LOGITS
     synonymous
    0.95
     belonging
    0.92
     unbeat
    0.85
     credible
    0.83
     unfit
    0.83
     indispensable
    0.82
     embod
    0.81
     illegitimate
    0.80
     unethical
    0.79
     unworthy
    0.78
    Activations Density 1.001%

    No Known Activations