© 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. 479
    Prev
    Next
    INDEX
    Explanations

    highlight bold terms in a text, particularly football-related content

    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
    410
    +0.13
    0.6%
    222
    +0.11
    0.5%
    741
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    762
    +0.13
    0.10
    168
    +0.11
    0.07
    718
    +0.10
    0.07
    Negative Logits
     Lug
    -0.67
     Meier
    -0.65
     portfolio
    -0.62
    ohyd
    -0.60
     Locke
    -0.59
     CV
    -0.59
    oku
    -0.58
    iste
    -0.57
     Malta
    -0.57
     Comprehensive
    -0.57
    POSITIVE LOGITS
     clam
    1.07
     applaud
    1.04
     alike
    1.02
     admire
    1.00
     disbel
    0.99
     ridicule
    0.97
     pity
    0.95
     cheering
    0.95
     envy
    0.94
     scorn
    0.93
    Activations Density 1.088%

    No Known Activations