© 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 24576
    9. 8-RES_FS24576-JB
    11. 7268
    Prev
    Next
    INDEX
    Explanations

    violent or threatening actions

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    343
    +0.13
    0.6%
    572
    +0.12
    0.5%
    732
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    343
    +0.13
    0.04
    362
    +0.12
    0.04
    572
    +0.11
    0.04
    Negative Logits
     pairing
    -0.86
    gradation
    -0.75
    omen
    -0.73
    ansion
    -0.72
    inctions
    -0.71
    RGB
    -0.71
    cation
    -0.70
    ownt
    -0.69
    ablishment
    -0.69
     conversions
    -0.69
    POSITIVE LOGITS
     knives
    1.23
     scissors
    1.20
     torches
    1.19
     pistols
    1.12
     guns
    1.08
     handguns
    1.08
     rifles
    1.07
     clipboard
    1.05
     knife
    1.03
     clenched
    1.01
    Activations Density 0.197%

    No Known Activations