© 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. 2905
    Prev
    Next
    INDEX
    Explanations

    provinces and political news related to energy and government decisions

    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
    538
    +0.13
    0.6%
    276
    +0.12
    0.5%
    624
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    538
    +0.13
    0.01
    326
    +0.12
    0.01
    288
    +0.11
    0.02
    Negative Logits
    aiman
    -0.77
    akening
    -0.77
     Mehran
    -0.75
    ourt
    -0.72
    oret
    -0.72
    enary
    -0.72
    atform
    -0.71
    estic
    -0.68
    imated
    -0.65
    duino
    -0.65
    POSITIVE LOGITS
    gaard
    1.13
    xual
    0.91
    lich
    0.90
    belt
    0.90
    gence
    0.89
    ance
    0.88
    lein
    0.84
    backer
    0.83
     Singh
    0.83
    holm
    0.82
    Activations Density 0.044%

    No Known Activations