© 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. 1356
    Prev
    Next
    INDEX
    Explanations

    references to political and philosophical concepts, specifically focusing on principles and values

    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
    394
    +0.11
    0.5%
    401
    +0.11
    0.5%
    288
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    401
    +0.11
    0.04
    247
    +0.11
    0.03
    255
    +0.11
    0.04
    Negative Logits
    otte
    -0.74
    noon
    -0.72
    sites
    -0.72
    charges
    -0.69
    ilation
    -0.69
    onga
    -0.68
    DERR
    -0.68
    fac
    -0.67
    azar
    -0.67
    aughter
    -0.67
    POSITIVE LOGITS
     principles
    1.15
     tenets
    1.04
     ideals
    1.03
     beliefs
    0.97
     embodied
    0.95
     guiding
    0.94
     precept
    0.94
     underpin
    0.93
     Principles
    0.92
     creed
    0.92
    Activations Density 0.177%

    No Known Activations