© 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 EvalsExportsinterp-engineNEWAPI 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. 4396
    Prev
    Next
    INDEX
    Explanations

    instances where the word 'back' is mentioned in relation to different contexts, such as positioning or turning

    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
    532
    +0.09
    0.4%
    541
    +0.09
    0.4%
    650
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    532
    +0.09
    0.03
    435
    +0.09
    0.03
    538
    +0.08
    0.03
    Negative Logits
    arenthood
    -0.68
     Recomm
    -0.64
    Adds
    -0.63
     rarity
    -0.58
    anto
    -0.58
     Arpaio
    -0.57
     bonding
    -0.57
     introduces
    -0.56
     interaction
    -0.56
    urdue
    -0.56
    POSITIVE LOGITS
     crossed
    0.76
     blown
    0.74
     burned
    0.74
     firmly
    0.73
    stretched
    0.73
     glued
    0.71
     locked
    0.70
     plugged
    0.69
     dirty
    0.69
     knees
    0.68
    Activations Density 0.096%

    No Known Activations