© 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. 7347
    Prev
    Next
    INDEX
    Explanations

    mentions of specific dates or days

    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
    430
    +0.14
    0.6%
    384
    +0.12
    0.6%
    105
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    430
    +0.14
    0.03
    384
    +0.12
    0.03
    105
    +0.11
    0.02
    Negative Logits
    eers
    -0.93
    eer
    -0.87
     Conquer
    -0.83
    OHN
    -0.76
    iframe
    -0.74
    ahime
    -0.72
    hips
    -0.68
    egal
    -0.66
    Control
    -0.66
    lasses
    -0.65
    POSITIVE LOGITS
     afternoon
    1.65
     morning
    1.54
     evening
    1.48
     night
    1.28
     mornings
    1.15
    morning
    1.11
     Evening
    0.95
     Night
    0.94
    night
    0.92
     Morning
    0.90
    Activations Density 0.025%

    No Known Activations