© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 49152
    9. 8-RES_FS49152-JB
    11. 43026
    Prev
    Next
    INDEX
    Explanations

    references to rights and duties, particularly in the context of civic responsibilities and legal frameworks

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    346
    +0.10
    0.4%
    595
    +0.10
    0.4%
    297
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    601
    +0.10
    0.01
    318
    +0.10
    0.01
    92
    +0.10
    0.01
    Negative Logits
     Sakuya
    -0.74
     Flo
    -0.70
     Vaugh
    -0.67
    bang
    -0.64
     Prescott
    -0.63
    kson
    -0.62
     documentaries
    -0.62
     Hai
    -0.62
     Rise
    -0.62
     sinks
    -0.60
    POSITIVE LOGITS
     discretion
    1.10
     patience
    1.09
     rights
    0.97
     privilege
    0.94
     utmost
    0.93
     autonomy
    0.92
     obligation
    0.91
     privileges
    0.89
     judgement
    0.87
     judgment
    0.87
    Activations Density 0.026%

    No Known Activations