© 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 12288
    9. 8-RES_FS12288-JB
    11. 714
    Prev
    Next
    INDEX
    Explanations

    technical guidance within a written document

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    377
    +0.10
    0.5%
    606
    +0.10
    0.5%
    221
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    377
    +0.10
    0.08
    490
    +0.10
    0.06
    271
    +0.10
    0.05
    Negative Logits
    ulner
    -0.66
    roots
    -0.63
    ŃĶ
    -0.61
    riots
    -0.59
    severe
    -0.59
    ãĥ©ãĥ³
    -0.58
    iann
    -0.57
    ãĤ¸
    -0.57
    ocom
    -0.57
    kefeller
    -0.56
    POSITIVE LOGITS
     but
    1.12
     though
    1.03
     tho
    1.02
     haha
    0.99
     anymore
    0.97
     BUT
    0.89
     :(
    0.87
     yet
    0.83
    but
    0.82
     But
    0.81
    Activations Density 0.868%

    No Known Activations