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 49152
    5. 8-RES_FS49152-JB
    6. 12041
    Prev
    Next
    INDEX
    Explanations

    references to political leaders and positions in governance

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    224
    +0.11
    0.5%
    687
    +0.10
    0.5%
    141
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    224
    +0.11
    0.05
    371
    +0.10
    0.04
    146
    +0.10
    0.04
    Negative Logits
    MpServer
    -0.74
    rawdownloadcloneembedreportprint
    -0.70
    maxwell
    -0.69
    ::::::::
    -0.69
    actory
    -0.66
    essen
    -0.65
    ASED
    -0.64
     fireplace
    -0.63
    ADD
    -0.63
    atlantic
    -0.63
    POSITIVE LOGITS
     Mahmoud
    1.19
     Bashar
    1.09
     Francois
    1.03
     Mahm
    1.00
     François
    1.00
     Hassan
    0.97
     Recep
    0.96
     Tayyip
    0.95
     Jinping
    0.95
    ial
    0.90
    Activations Density 0.052%

    No Known Activations