© 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. 17148
    Prev
    Next
    INDEX
    Explanations

    mentions of the location "Alton"

    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
    128
    +0.14
    0.6%
    515
    +0.10
    0.4%
    414
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    122
    +0.14
    0.02
    232
    +0.10
    0.02
    630
    +0.10
    0.02
    Negative Logits
    saf
    -0.71
    ptive
    -0.71
    PER
    -0.70
     JPM
    -0.68
    glers
    -0.67
    ulative
    -0.65
    ptions
    -0.64
    stract
    -0.63
    Magikarp
    -0.63
    BILITIES
    -0.63
    POSITIVE LOGITS
    neau
    0.95
    nel
    0.92
    ews
    0.89
    nian
    0.89
    nia
    0.86
    icum
    0.85
    ©¶æ¥µ
    0.81
    osaurus
    0.80
    ians
    0.78
    ium
    0.77
    Activations Density 0.055%

    No Known Activations