© 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. 17353
    Prev
    Next
    INDEX
    Explanations

    phrases related to criticism or negative portrayal

    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
    260
    +0.11
    0.5%
    70
    +0.10
    0.5%
    186
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    70
    +0.11
    0.04
    486
    +0.10
    0.02
    665
    +0.10
    0.03
    Negative Logits
    ItemTracker
    -0.72
    PsyNetMessage
    -0.70
     Surv
    -0.64
    Avg
    -0.61
     Storage
    -0.59
     railing
    -0.59
     forbids
    -0.59
     Advice
    -0.59
     Variant
    -0.58
     summar
    -0.57
    POSITIVE LOGITS
    vind
    0.90
     smarter
    0.84
     stronger
    0.80
    able
    0.79
    uate
    0.74
    ingly
    0.71
     louder
    0.71
     worse
    0.70
    thood
    0.70
    ables
    0.70
    Activations Density 0.081%

    No Known Activations