© 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. 14314
    Prev
    Next
    INDEX
    Explanations

    phrases related to heavy physical objects or objects that require physical strength

    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
    590
    +0.10
    0.5%
    666
    +0.10
    0.5%
    571
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    590
    +0.10
    0.02
    666
    +0.10
    0.02
    65
    +0.10
    0.02
    Negative Logits
    icho
    -0.86
    DragonMagazine
    -0.82
    urated
    -0.76
    Ĥİ
    -0.74
    çīĪ
    -0.74
     Unlimited
    -0.73
    sure
    -0.73
    DEV
    -0.71
    uably
    -0.71
     Kirin
    -0.70
    POSITIVE LOGITS
    weights
    1.71
     hitters
    1.27
    weight
    1.25
     hitter
    1.18
     lifting
    1.15
     weights
    1.05
     duty
    1.02
     weight
    0.99
    metal
    0.96
     workload
    0.95
    Activations Density 0.029%

    No Known Activations