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 24576
    5. 8-RES_FS24576-JB
    6. 314
    Prev
    Next
    INDEX
    Explanations

    phrases related to personal or organizational historical actions or involvement

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    377
    +0.11
    0.5%
    390
    +0.11
    0.5%
    422
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    390
    +0.11
    0.02
    724
    +0.11
    0.03
    575
    +0.09
    0.02
    Negative Logits
    wered
    -0.75
    ando
    -0.71
    ower
    -0.70
    hap
    -0.68
    ople
    -0.68
    idden
    -0.66
    oner
    -0.65
    ighth
    -0.64
    uri
    -0.63
    agers
    -0.63
    POSITIVE LOGITS
     dating
    0.86
     associ
    0.78
    acqu
    0.76
     revolving
    0.74
     dealings
    0.74
     precedent
    0.72
     relationship
    0.72
     fraught
    0.72
     history
    0.72
     affiliation
    0.71
    Activations Density 0.054%

    No Known Activations