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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 41236
    Prev
    Next
    INDEX
    Explanations

    words and phrases associated with political actions and statements

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    86
    +0.09
    0.4%
    546
    +0.08
    0.4%
    226
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    546
    +0.09
    0.07
    43
    +0.08
    0.06
    720
    +0.08
    0.06
    Negative Logits
    iple
    -0.96
    raft
    -0.65
     Cros
    -0.64
    atars
    -0.63
    uckland
    -0.63
    utic
    -0.62
     Shard
    -0.62
    uber
    -0.61
     Hatt
    -0.61
     Cube
    -0.60
    POSITIVE LOGITS
     thereby
    1.11
     including
    0.97
     whereas
    0.97
     perpetrated
    0.96
     despite
    0.95
     citing
    0.95
    while
    0.95
     aka
    0.94
     which
    0.94
     lest
    0.93
    Activations Density 0.296%

    No Known Activations