© 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. 1246
    Prev
    Next
    INDEX
    Explanations

    the word "have" followed by either a high activation word or a number

    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
    68
    +0.12
    0.5%
    280
    +0.09
    0.4%
    547
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    566
    +0.12
    0.04
    344
    +0.09
    0.04
    405
    +0.09
    0.04
    Negative Logits
    anding
    -0.65
    atin
    -0.61
    catentry
    -0.60
    osi
    -0.58
     Crown
    -0.57
    wait
    -0.57
     plaintiff
    -0.57
    POS
    -0.56
    etting
    -0.56
    amen
    -0.56
    POSITIVE LOGITS
     been
    1.32
     gotten
    1.21
    been
    1.18
     undergone
    1.06
     Been
    1.04
    gotten
    1.01
     gone
    0.97
     begun
    0.94
     fallen
    0.92
     arisen
    0.92
    Activations Density 0.140%

    No Known Activations