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 1536
    5. 8-RES_FS1536-JB
    6. 897
    Prev
    Next
    INDEX
    Explanations

    verbs related to actions and activities in a specific context

    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
    447
    +0.16
    0.7%
    70
    +0.11
    0.5%
    7
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    70
    +0.16
    0.18
    665
    +0.11
    0.17
    570
    +0.11
    0.06
    Negative Logits
    fal
    -0.66
    hon
    -0.65
    Enlarge
    -0.64
    osi
    -0.64
    icus
    -0.62
    rin
    -0.58
    pora
    -0.58
    quart
    -0.58
    nton
    -0.58
    enei
    -0.58
    POSITIVE LOGITS
     anything
    1.05
     them
    0.99
     everything
    0.89
     something
    0.88
     freely
    0.87
     differently
    0.87
     alot
    0.85
     oneself
    0.82
     him
    0.78
     somebody
    0.78
    Activations Density 3.324%

    No Known Activations