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. 23277
    Prev
    Next
    INDEX
    Explanations

    terms related to computational concepts and methodologies

    oai_token-act-pair · gpt-4o-miniTriggered 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
    481
    +0.32
    1.6%
    447
    +0.24
    1.2%
    67
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.32
    0.10
    67
    +0.24
    0.08
    481
    +0.10
    0.08
    Negative Logits
    swer
    -0.67
    clave
    -0.64
     Pwr
    -0.63
     Died
    -0.62
     suite
    -0.62
    zzle
    -0.60
     loader
    -0.59
    fecture
    -0.59
     fielder
    -0.58
    bsite
    -0.58
    POSITIVE LOGITS
    Semitism
    0.85
    paren
    0.77
     warfare
    0.76
    ism
    0.76
    ivism
    0.75
     neutrality
    0.75
     capitalism
    0.73
     feminism
    0.73
     weddings
    0.72
     parenting
    0.71
    Activations Density 0.705%

    No Known Activations