© 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. 9715
    Prev
    Next
    INDEX
    Explanations

    direct questions in quoted dialogue

    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
    255
    +0.12
    0.6%
    186
    +0.10
    0.5%
    6
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.12
    0.07
    255
    +0.10
    0.05
    426
    +0.10
    0.04
    Negative Logits
    ãĥķãĤ¡
    -0.68
    unker
    -0.67
    ciplinary
    -0.66
    aser
    -0.65
    ancies
    -0.64
    ayan
    -0.63
    Reviewer
    -0.63
    asers
    -0.63
    sels
    -0.63
    central
    -0.62
    POSITIVE LOGITS
    /"
    0.95
     whereas
    0.88
     Then
    0.88
    [/
    0.83
     etc
    0.80
     followed
    0.79
     Similarly
    0.77
     Tears
    0.75
     referring
    0.75
     haha
    0.74
    Activations Density 0.313%

    No Known Activations