© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 12288
    9. 8-RES_FS12288-JB
    11. 10186
    Prev
    Next
    INDEX
    Explanations

    word parts related to historical or religious figures and ceremonies

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    +0.12
    0.5%
    496
    +0.10
    0.5%
    738
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.12
    0.03
    232
    +0.10
    0.03
    738
    +0.10
    0.02
    Negative Logits
    ivia
    -0.78
     Oy
    -0.75
    ores
    -0.69
     Tasman
    -0.67
    estation
    -0.67
     Samoa
    -0.66
    eenth
    -0.66
    ļé
    -0.65
     Hots
    -0.65
    ured
    -0.64
    POSITIVE LOGITS
    rha
    1.03
    abbit
    0.98
    rr
    0.96
    lich
    0.94
    ansom
    0.89
    acket
    0.85
    andom
    0.83
    ror
    0.82
    anger
    0.81
    aternity
    0.81
    Activations Density 0.153%

    No Known Activations