© 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. 9653
    Prev
    Next
    INDEX
    Explanations

    the word "first"

    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
    373
    +0.14
    0.7%
    242
    +0.10
    0.5%
    530
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    366
    +0.14
    0.03
    242
    +0.10
    0.04
    530
    +0.09
    0.03
    Negative Logits
    peer
    -0.71
    rien
    -0.70
    Max
    -0.65
    rend
    -0.65
    roth
    -0.64
    meyer
    -0.64
    lov
    -0.64
    bos
    -0.62
    suits
    -0.60
    styles
    -0.60
    POSITIVE LOGITS
     sake
    1.01
     installment
    0.92
     consecutive
    0.82
     longest
    0.82
     decade
    0.80
     eternity
    0.80
     instance
    0.79
     purposes
    0.79
     millennium
    0.77
     reasons
    0.77
    Activations Density 0.076%

    No Known Activations