© 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 49152
    9. 8-RES_FS49152-JB
    11. 37280
    Prev
    Next
    INDEX
    Explanations

    phrases indicating priorities or commitments

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    608
    +0.13
    0.6%
    216
    +0.11
    0.5%
    88
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    608
    +0.13
    0.05
    88
    +0.11
    0.03
    22
    +0.10
    0.03
    Negative Logits
    soDeliveryDate
    -0.90
    nick
    -0.76
    cles
    -0.73
    qi
    -0.70
    é¾
    -0.70
    ãĤ´ãĥ³
    -0.69
    aunder
    -0.69
    opus
    -0.67
    mol
    -0.67
    loo
    -0.67
    POSITIVE LOGITS
     policymakers
    1.07
    geries
    0.99
     purposes
    0.96
     determining
    0.93
    gery
    0.92
     planners
    0.92
    bidden
    0.88
     employers
    0.83
     us
    0.82
     generations
    0.82
    Activations Density 0.104%

    No Known Activations