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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 24576
    5. 8-RES_FS24576-JB
    6. 280
    Prev
    Next
    INDEX
    Explanations

    statements starting with "There is"

    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
    20
    +0.12
    0.6%
    455
    +0.10
    0.5%
    674
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    20
    +0.12
    0.07
    398
    +0.10
    0.06
    126
    +0.10
    0.05
    Negative Logits
    ares
    -0.83
    bies
    -0.75
    wake
    -0.63
    ends
    -0.63
     cylinders
    -0.62
     retrie
    -0.60
    ooters
    -0.59
    letes
    -0.58
     closes
    -0.57
     wastes
    -0.57
    POSITIVE LOGITS
     plenty
    1.01
     ample
    0.96
     no
    0.92
     always
    0.91
    女
    0.88
     definitely
    0.84
     currently
    0.83
     precedent
    0.83
     nothing
    0.81
     ALWAYS
    0.80
    Activations Density 0.110%

    No Known Activations