© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 19814
    Prev
    Next
    INDEX
    Explanations

    social media-related terms like "Facebook," "Twitter," and "Google."

    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
    613
    +0.12
    0.6%
    136
    +0.12
    0.6%
    218
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    74
    +0.12
    0.02
    120
    +0.12
    0.02
    136
    +0.10
    0.02
    Negative Logits
    ģĸ
    -0.67
    onics
    -0.65
    iless
    -0.63
     recomp
    -0.63
    ictional
    -0.62
     fres
    -0.61
    enza
    -0.61
    iqueness
    -0.61
     forthcoming
    -0.61
    :]
    -0.59
    POSITIVE LOGITS
    edin
    0.94
    facebook
    0.86
    LinkedIn
    0.85
    Facebook
    0.81
    Reddit
    0.80
    Apps
    0.79
     Whats
    0.77
     Instagram
    0.77
     hashtag
    0.76
     LinkedIn
    0.76
    Activations Density 0.079%

    No Known Activations