© 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. 626
    Prev
    Next
    INDEX
    Explanations

    phrases related to the importance of specific actions or concepts

    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
    285
    +0.11
    0.5%
    618
    +0.10
    0.4%
    511
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    285
    +0.11
    0.03
    511
    +0.10
    0.03
    737
    +0.09
    0.03
    Negative Logits
     Wonderland
    -0.72
    wagen
    -0.69
    dit
    -0.65
    cker
    -0.61
    icut
    -0.58
    upt
    -0.58
    Brend
    -0.58
    rooms
    -0.57
     Emirates
    -0.57
     Sheep
    -0.56
    POSITIVE LOGITS
    lessness
    1.12
     tremend
    0.97
     importance
    0.93
    uality
    0.92
    iness
    0.89
     inherent
    0.86
    worthiness
    0.77
     factor
    0.76
    lessly
    0.74
    fulness
    0.73
    Activations Density 0.101%

    No Known Activations