© 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. 871
    Prev
    Next
    INDEX
    Explanations

    phrases related to controversial or serious topics

    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
    377
    +0.12
    0.5%
    613
    +0.11
    0.5%
    246
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    151
    +0.12
    0.03
    377
    +0.11
    0.03
    0
    +0.10
    0.02
    Negative Logits
    ifully
    -0.82
    anic
    -0.76
    tein
    -0.73
    atu
    -0.73
    winner
    -0.72
    orius
    -0.71
    sylv
    -0.69
     Parenthood
    -0.69
    anooga
    -0.67
    çİĭ
    -0.66
    POSITIVE LOGITS
     dent
    0.81
     retarded
    0.76
     entertained
    0.76
     injuring
    0.76
     ill
    0.76
     considered
    0.73
     injure
    0.72
     doubted
    0.72
     challenged
    0.70
     considering
    0.70
    Activations Density 0.057%

    No Known Activations