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. 23385
    Prev
    Next
    INDEX
    Explanations

    the phrase "away" and its various contexts of distance or separation

    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
    696
    +0.10
    0.5%
    615
    +0.09
    0.4%
    537
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    589
    +0.10
    0.02
    565
    +0.09
    0.02
    222
    +0.09
    0.02
    Negative Logits
    nce
    -0.70
    ificial
    -0.69
    imar
    -0.65
    inse
    -0.65
    oleon
    -0.64
    uv
    -0.63
    =-=-=-=-=-=-=-=-
    -0.61
    cess
    -0.61
    oly
    -0.60
    isms
    -0.60
    POSITIVE LOGITS
    aneously
    0.76
    lier
    0.73
    sighted
    0.71
     from
    0.70
    angs
    0.67
     closest
    0.67
     اÙĦ
    0.62
     ahead
    0.61
     earthqu
    0.61
     surpass
    0.59
    Activations Density 0.051%

    No Known Activations