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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 475
    Prev
    Next
    INDEX
    Explanations

    dialogue including profanity and threats

    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
    572
    +0.12
    0.6%
    279
    +0.12
    0.5%
    683
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    279
    +0.12
    0.11
    572
    +0.12
    0.11
    331
    +0.10
    0.10
    Negative Logits
    etheless
    -0.85
    surprisingly
    -0.82
     anecd
    -0.80
     unsurprisingly
    -0.73
     largely
    -0.72
     predominantly
    -0.70
     predictably
    -0.69
    xtap
    -0.68
    uably
    -0.68
    prisingly
    -0.68
    POSITIVE LOGITS
     â̦"
    1.48
    !'"
    1.43
    â̦"
    1.33
    !"
    1.32
    ',"
    1.31
    ..."
    1.29
    ,'"
    1.27
    '"
    1.26
    !",
    1.26
    '."
    1.22
    Activations Density 3.726%

    No Known Activations