Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 12288
    5. 8-RES_FS12288-JB
    6. 23
    Prev
    Next
    INDEX
    Explanations

    words related to authority and permission

    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
    498
    +0.13
    0.6%
    642
    +0.11
    0.5%
    59
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    75
    +0.13
    0.04
    498
    +0.11
    0.04
    403
    +0.10
    0.03
    Negative Logits
    xon
    -0.83
    borough
    -0.81
    bard
    -0.78
    kes
    -0.72
    enegger
    -0.71
    bage
    -0.69
    inger
    -0.67
    nard
    -0.66
    econom
    -0.64
     redes
    -0.64
    POSITIVE LOGITS
     exceptions
    0.80
     us
    0.77
     exemptions
    0.77
     unrestricted
    0.75
    Reviewer
    0.72
     deviations
    0.69
     deductions
    0.69
     him
    0.68
     them
    0.68
     visitors
    0.67
    Activations Density 0.111%

    No Known Activations