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

    Neuronpedia

    APISteerSAE EvalsBlog/PodcastNEWSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlog/PodcastGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 3072
    5. 8-RES_FS3072-JB
    6. 2824
    Prev
    Next
    INDEX
    Explanations

    sentences discussing financial or economic matters

    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
    679
    +0.20
    0.9%
    447
    +0.15
    0.7%
    640
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.20
    0.23
    679
    +0.15
    0.21
    440
    +0.13
    0.16
    Negative Logits
    ascus
    -0.74
     gravy
    -0.73
     unob
    -0.73
     warr
    -0.71
     unus
    -0.71
     minim
    -0.71
     whis
    -0.71
     maniac
    -0.70
     optional
    -0.70
     alley
    -0.69
    POSITIVE LOGITS
     Firstly
    1.74
     Whereas
    1.54
     Specifically
    1.47
     Especially
    1.46
     Particularly
    1.41
     According
    1.38
     Suppose
    1.35
     Instead
    1.33
     Though
    1.30
     Whether
    1.29
    Activations Density 1.767%

    No Known Activations