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 1536
    5. 8-RES_FS1536-JB
    6. 420
    Prev
    Next
    INDEX
    Explanations

    mentions of death and related events

    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
    223
    +0.13
    0.6%
    759
    +0.12
    0.6%
    319
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    759
    +0.13
    0.10
    223
    +0.12
    0.09
    319
    +0.11
    0.08
    Negative Logits
    MN
    -0.74
    umar
    -0.73
     Hilbert
    -0.66
     Libre
    -0.65
    Cola
    -0.65
    orney
    -0.65
     CLSID
    -0.64
    OPER
    -0.63
     Lank
    -0.63
    Americ
    -0.62
    POSITIVE LOGITS
     reckoning
    0.90
    locked
    0.90
     toll
    0.85
    adder
    0.84
     tragically
    0.84
    blow
    0.83
    heading
    0.80
     deaths
    0.78
    thirst
    0.77
    switch
    0.77
    Activations Density 1.691%

    No Known Activations