Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 1536
    5. 8-RES_FS1536-JB
    6. 965
    Prev
    Next
    INDEX
    Explanations

    phrases related to offering information or providing guidance

    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
    634
    +0.12
    0.5%
    431
    +0.10
    0.5%
    566
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    566
    +0.12
    0.14
    718
    +0.10
    0.12
    634
    +0.10
    0.12
    Negative Logits
     Farming
    -0.64
    arthed
    -0.62
     Globe
    -0.61
    ource
    -0.60
     Feast
    -0.59
     reportedly
    -0.58
    restling
    -0.57
     Regions
    -0.57
     Gul
    -0.57
    è¦ļéĨĴ
    -0.57
    POSITIVE LOGITS
     deserved
    0.91
     unfairly
    0.81
     misunderstood
    0.81
     underestimated
    0.79
     underrated
    0.78
     deserve
    0.78
     fit
    0.77
     ought
    0.77
     deserves
    0.76
    should
    0.74
    Activations Density 3.046%

    No Known Activations