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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 26193
    Prev
    Next
    INDEX
    Explanations

    expressions of belief and declarations of love

    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
    640
    +0.28
    1.3%
    525
    +0.10
    0.4%
    143
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.28
    0.03
    6
    +0.10
    0.02
    160
    +0.09
    0.02
    Negative Logits
    etheless
    -0.71
     manoeuv
    -0.70
    hement
    -0.69
     thrust
    -0.68
     aiming
    -0.68
     footing
    -0.63
     apprehended
    -0.63
    arers
    -0.61
     either
    -0.60
     ultr
    -0.60
    POSITIVE LOGITS
    ³³³³
    0.97
    ³³³
    0.93
    CHAPTER
    0.90
    Bonus
    0.87
    EDIT
    0.85
    Original
    0.84
    BOOK
    0.80
    ccording
    0.79
    Copyright
    0.78
    QUEST
    0.78
    Activations Density 0.028%

    No Known Activations