Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 24576
    5. 8-RES_FS24576-JB
    6. 11191
    Prev
    Next
    INDEX
    Explanations

    words related to controversial social or political topics

    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
    357
    +0.10
    0.4%
    449
    +0.09
    0.4%
    99
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    748
    +0.10
    0.03
    312
    +0.09
    0.03
    28
    +0.09
    0.03
    Negative Logits
    igger
    -0.84
    imet
    -0.76
    ente
    -0.74
    agger
    -0.73
    utenberg
    -0.73
    oused
    -0.73
    actionDate
    -0.72
    imeters
    -0.72
    ĸļ
    -0.72
    unes
    -0.71
    POSITIVE LOGITS
     importantly
    0.86
     thereafter
    0.78
     secondly
    0.74
     eliminate
    0.74
     starve
    0.72
     guessed
    0.71
     succumb
    0.71
     vice
    0.71
     eliminates
    0.70
     consequently
    0.69
    Activations Density 0.118%

    No Known Activations