© 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. 20690
    Prev
    Next
    INDEX
    Explanations

    verbs or nouns related to negative or harmful actions

    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
    35
    +0.12
    0.6%
    326
    +0.11
    0.5%
    269
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.12
    0.01
    269
    +0.11
    0.01
    61
    +0.11
    0.01
    Negative Logits
     reluct
    -0.68
    duino
    -0.67
    avior
    -0.67
    arij
    -0.66
    WER
    -0.65
    DragonMagazine
    -0.65
    taboola
    -0.64
     compressor
    -0.63
    rict
    -0.62
    leground
    -0.62
    POSITIVE LOGITS
     havoc
    1.26
    ous
    0.96
    ously
    0.88
    OUS
    0.86
    ifully
    0.83
     spoil
    0.81
    iful
    0.76
    nit
    0.74
     lives
    0.72
     ruining
    0.71
    Activations Density 0.037%

    No Known Activations