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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 1536
    5. 8-RES_FS1536-JB
    6. 875
    Prev
    Next
    INDEX
    Explanations

    discussions related to social, political, and economic issues

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs1536-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    1,536
    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
    373
    +0.13
    0.6%
    39
    +0.11
    0.5%
    101
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    739
    +0.13
    0.05
    39
    +0.11
    0.09
    306
    +0.10
    0.08
    Negative Logits
    aunts
    -0.79
    Tokens
    -0.68
     manif
    -0.65
     millionaires
    -0.65
    rums
    -0.64
     woes
    -0.63
    trap
    -0.63
    gap
    -0.62
    hell
    -0.61
    zza
    -0.61
    POSITIVE LOGITS
     copyrighted
    1.06
     copyright
    1.03
     trademarks
    0.92
     Copyright
    0.89
     reprint
    0.88
     editorial
    0.87
    Disclaimer
    0.86
     disclaimer
    0.85
     formatting
    0.85
     affiliate
    0.85
    Activations Density 3.110%

    No Known Activations