© 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. 668
    Prev
    Next
    INDEX
    Explanations

    mentions of sports plays and player performance

    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
    761
    +0.12
    0.5%
    216
    +0.10
    0.5%
    540
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    761
    +0.12
    0.05
    540
    +0.10
    0.04
    271
    +0.10
    0.04
    Negative Logits
    brow
    -0.66
    İ
    -0.65
     Shipping
    -0.65
     Beir
    -0.63
    fty
    -0.62
    icon
    -0.62
    imb
    -0.61
    Ĭ±
    -0.58
    appro
    -0.57
    nda
    -0.57
    POSITIVE LOGITS
    wright
    0.96
    ername
    0.91
     alongside
    0.88
     colleg
    0.86
     against
    0.85
     played
    0.76
    imony
    0.74
     poker
    0.74
    halla
    0.74
     professionally
    0.71
    Activations Density 0.084%

    No Known Activations