Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 1536
    5. 8-RES_FS1536-JB
    6. 1268
    Prev
    Next
    INDEX
    Explanations

    repetitive use of the phrase "a lot" followed by a number in textual content

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs1536-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    1,536
    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
    71
    +0.14
    0.7%
    596
    +0.11
    0.5%
    241
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    344
    +0.14
    0.08
    596
    +0.11
    0.08
    71
    +0.10
    0.06
    Negative Logits
    heid
    -0.77
    Constructed
    -0.72
    ufact
    -0.65
    ESCO
    -0.65
    agate
    -0.63
    perty
    -0.63
     sole
    -0.61
     thence
    -0.59
     slightest
    -0.57
    sburgh
    -0.57
    POSITIVE LOGITS
     of
    0.78
     nicer
    0.69
    bos
    0.67
    ðŁij
    0.67
    buck
    0.67
    rage
    0.66
    tery
    0.65
    nz
    0.65
    picture
    0.63
    UID
    0.62
    Activations Density 0.896%

    No Known Activations