© 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
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 12288
    5. 8-RES_FS12288-JB
    6. 315
    Prev
    Next
    INDEX
    Explanations

    phrases related to failure or inadequacy

    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
    665
    +0.17
    0.8%
    718
    +0.15
    0.7%
    762
    +0.14
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.17
    0.05
    718
    +0.15
    0.05
    49
    +0.14
    0.04
    Negative Logits
    soDeliveryDate
    -0.91
    calling
    -0.73
    mails
    -0.70
    Reader
    -0.70
     Printed
    -0.69
    rolled
    -0.67
     Used
    -0.66
     Caller
    -0.66
    umes
    -0.65
    soType
    -0.65
    POSITIVE LOGITS
     comply
    1.37
     qualify
    1.25
     satisfy
    1.24
     adequately
    1.23
     resolve
    1.23
     heed
    1.23
     fulfill
    1.22
     adhere
    1.21
     recognize
    1.21
     grasp
    1.18
    Activations Density 0.106%

    No Known Activations