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 3072
    5. 8-RES_FS3072-JB
    6. 2945
    Prev
    Next
    INDEX
    Explanations

    words related to textual formatting and software tools

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    564
    +0.12
    0.5%
    74
    +0.11
    0.5%
    309
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    762
    +0.12
    0.12
    564
    +0.11
    0.09
    575
    +0.09
    0.10
    Negative Logits
     festive
    -0.69
     jaw
    -0.67
    xtap
    -0.64
    ãĥ©ãĥ³
    -0.63
     sugg
    -0.61
     Kul
    -0.61
    ittees
    -0.59
     Top
    -0.57
     Phill
    -0.56
     Kaf
    -0.56
    POSITIVE LOGITS
     itself
    0.84
     nonetheless
    0.83
     anyways
    0.81
     ain
    0.81
     oneself
    0.80
     anyway
    0.79
     Himself
    0.77
     ours
    0.77
     everywhere
    0.76
     minus
    0.74
    Activations Density 2.360%

    No Known Activations