© 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 768
    9. 8-RES_FS768-JB
    11. 711
    Prev
    Next
    INDEX
    Explanations

    phrases related to fighting or contesting

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    513
    +0.12
    0.5%
    247
    +0.11
    0.5%
    496
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.12
    0.18
    247
    +0.11
    0.11
    531
    +0.10
    0.11
    Negative Logits
    ocument
    -0.81
    essage
    -0.76
    arton
    -0.76
    ascript
    -0.71
    mberg
    -0.71
    uras
    -0.70
     Atmospheric
    -0.70
     Seym
    -0.67
    seys
    -0.66
    ICAN
    -0.65
    POSITIVE LOGITS
     against
    1.30
     waged
    1.03
     raged
    0.98
    against
    0.97
     fought
    0.94
     raging
    0.93
     Against
    0.93
     fiercely
    0.89
    front
    0.86
    fights
    0.86
    Activations Density 4.417%

    No Known Activations