© 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 49152
    5. 8-RES_FS49152-JB
    6. 37600
    Prev
    Next
    INDEX
    Explanations

    the letter 'k' in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    659
    +0.14
    0.7%
    593
    +0.12
    0.5%
    314
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.14
    0.02
    701
    +0.12
    0.02
    232
    +0.11
    0.02
    Negative Logits
     mosqu
    -0.68
     Mayweather
    -0.61
     traitor
    -0.60
     celebr
    -0.60
     derogatory
    -0.60
     kernels
    -0.60
     commons
    -0.58
     advers
    -0.58
     Staples
    -0.57
     panic
    -0.57
    POSITIVE LOGITS
    ull
    1.05
    rieg
    1.04
    irts
    1.02
    ansas
    1.02
    yrim
    1.02
    irk
    0.98
    arten
    0.94
    ngth
    0.92
    utz
    0.90
    rift
    0.89
    Activations Density 0.022%

    No Known Activations