© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 3725
    Prev
    Next
    INDEX
    Explanations

    phrases related to efforts or actions being futile or unsuccessful

    oai_token-act-pair · gpt-3.5-turbo
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    727
    +0.13
    0.6%
    34
    +0.11
    0.5%
    557
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    151
    +0.13
    0.03
    34
    +0.11
    0.02
    288
    +0.10
    0.02
    Negative Logits
    orthy
    -0.84
    APH
    -0.77
    origin
    -0.72
    andise
    -0.70
    ocard
    -0.70
    RN
    -0.66
    ainers
    -0.64
    hens
    -0.62
    avis
    -0.62
    eus
    -0.61
    POSITIVE LOGITS
     futile
    0.96
    ness
    0.91
    ly
    0.88
    nesses
    0.87
     attempts
    0.84
     attempt
    0.84
     miser
    0.83
    lust
    0.82
     pleas
    0.77
    cia
    0.76
    Activations Density 0.058%

    No Known Activations