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 49152
    5. 8-RES_FS49152-JB
    6. 41318
    Prev
    Next
    INDEX
    Explanations

    phrases indicating consequences or implications

    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
    267
    +0.11
    0.5%
    739
    +0.11
    0.5%
    483
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    8
    +0.11
    0.03
    239
    +0.11
    0.02
    688
    +0.10
    0.02
    Negative Logits
    oked
    -0.74
     redes
    -0.68
    uded
    -0.67
     accomp
    -0.66
    iazep
    -0.65
     withd
    -0.65
    kw
    -0.64
     sidx
    -0.64
     promot
    -0.64
    thia
    -0.63
    POSITIVE LOGITS
     goodbye
    0.91
     nothing
    0.76
     everything
    0.72
     fewer
    0.72
     that
    0.71
    little
    0.68
     whichever
    0.65
     exactly
    0.65
     we
    0.65
     they
    0.64
    Activations Density 0.025%

    No Known Activations