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. 23289
    Prev
    Next
    INDEX
    Explanations

    expressions of understanding or comprehension

    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
    750
    +0.10
    0.5%
    242
    +0.10
    0.5%
    521
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    750
    +0.10
    0.03
    521
    +0.10
    0.02
    715
    +0.10
    0.03
    Negative Logits
    mage
    -0.76
    estial
    -0.72
    gins
    -0.71
    thood
    -0.69
    aunder
    -0.69
    chieve
    -0.69
    idav
    -0.68
    ixt
    -0.67
    rouse
    -0.67
     Accessed
    -0.65
    POSITIVE LOGITS
     passionately
    0.73
     why
    0.73
    poke
    0.69
     firsthand
    0.69
     frustration
    0.69
     sadness
    0.69
     impat
    0.67
     personally
    0.67
     assurances
    0.67
     criticism
    0.66
    Activations Density 0.098%

    No Known Activations