© 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. 639
    Prev
    Next
    INDEX
    Explanations

    words related to corruption and unethical behavior

    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
    421
    +0.09
    0.4%
    733
    +0.09
    0.4%
    566
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    733
    +0.09
    0.05
    408
    +0.09
    0.09
    364
    +0.08
    0.06
    Negative Logits
    ellen
    -0.74
     Murd
    -0.72
    rick
    -0.70
    uay
    -0.68
    ntil
    -0.68
     Polo
    -0.66
    achus
    -0.66
    oil
    -0.65
    ysc
    -0.65
    Diamond
    -0.65
    POSITIVE LOGITS
    ties
    0.89
    ities
    0.85
     mundane
    0.72
     shenan
    0.69
     phenomena
    0.68
     drastic
    0.68
    worldly
    0.68
     specificity
    0.66
    complex
    0.66
     gems
    0.64
    Activations Density 2.933%

    No Known Activations