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. 179
    Prev
    Next
    INDEX
    Explanations

    verbs related to communication and interaction

    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
    725
    +0.11
    0.5%
    447
    +0.11
    0.5%
    415
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    274
    +0.11
    0.11
    195
    +0.11
    0.06
    642
    +0.10
    0.09
    Negative Logits
    onew
    -0.69
    abases
    -0.69
    enium
    -0.69
    kie
    -0.68
     ILCS
    -0.67
    quer
    -0.66
    agos
    -0.66
    onz
    -0.65
    tes
    -0.65
    brance
    -0.64
    POSITIVE LOGITS
     how
    0.99
     aloud
    0.85
     inconsistencies
    0.85
     anew
    0.83
     similarities
    0.82
     unequivocally
    0.79
     shortcomings
    0.76
     why
    0.76
     deficiencies
    0.75
     irregularities
    0.74
    Activations Density 1.757%

    No Known Activations