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. 46059
    Prev
    Next
    INDEX
    Explanations

    references to China and its geopolitical activities

    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
    481
    +0.14
    0.6%
    67
    +0.13
    0.6%
    635
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    67
    +0.14
    0.05
    461
    +0.13
    0.04
    57
    +0.12
    0.03
    Negative Logits
    schild
    -0.72
    RH
    -0.71
    âĢ¢âĢ¢âĢ¢âĢ¢
    -0.70
    veyard
    -0.68
    angled
    -0.68
     Payton
    -0.68
    iotic
    -0.67
    vind
    -0.67
    odox
    -0.67
    docker
    -0.66
    POSITIVE LOGITS
     Jinping
    1.12
     Unic
    1.03
     Yuan
    1.01
     Zhu
    0.93
     Jing
    0.91
     Xin
    0.90
     yuan
    0.87
     nationals
    0.86
     Guang
    0.84
     Telecom
    0.82
    Activations Density 0.028%

    No Known Activations