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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 10450
    Prev
    Next
    INDEX
    Explanations

    years expressed as 4-digit numbers corresponding to important events or historical facts

    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
    689
    +0.14
    0.6%
    390
    +0.11
    0.5%
    598
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    689
    +0.14
    0.05
    306
    +0.11
    0.05
    598
    +0.10
    0.03
    Negative Logits
     snipp
    -0.75
     pse
    -0.60
     Pru
    -0.60
     Hunts
    -0.58
    detail
    -0.57
    imore
    -0.57
     Jugg
    -0.57
    Dialogue
    -0.56
    umbn
    -0.56
    Jackson
    -0.55
    POSITIVE LOGITS
     onwards
    0.92
     onward
    0.80
    âĢ²
    0.73
    ties
    0.71
    quist
    0.70
    cards
    0.70
    staking
    0.63
     inception
    0.62
    cation
    0.61
    maxwell
    0.59
    Activations Density 0.098%

    No Known Activations