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 24576
    5. 8-RES_FS24576-JB
    6. 21629
    Prev
    Next
    INDEX
    Explanations

    mentions of specific laws or regulations

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    58
    +0.10
    0.5%
    636
    +0.10
    0.5%
    192
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.10
    0.07
    717
    +0.10
    0.03
    225
    +0.09
    0.04
    Negative Logits
    yssey
    -0.76
    ifted
    -0.69
     Rescue
    -0.63
    EStream
    -0.60
     ACTIONS
    -0.60
    ificent
    -0.59
    ilant
    -0.59
     Drawn
    -0.58
    CHO
    -0.57
    obl
    -0.57
    POSITIVE LOGITS
     altogether
    1.73
     entirely
    1.23
     obsolete
    1.01
     pesky
    0.96
     completely
    0.90
     redundant
    0.89
     outright
    0.89
     inhib
    0.87
     prematurely
    0.84
     superflu
    0.83
    Activations Density 0.479%

    No Known Activations