© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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 49152
    9. 8-RES_FS49152-JB
    11. 45767
    Prev
    Next
    INDEX
    Explanations

    negative phrases and expressions of refusal or rejection

    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
    359
    +0.09
    0.4%
    600
    +0.09
    0.4%
    107
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.09
    0.03
    579
    +0.09
    0.03
    687
    +0.09
    0.03
    Negative Logits
    effects
    -0.68
    unknown
    -0.65
    rawled
    -0.65
    Must
    -0.64
    DragonMagazine
    -0.64
     landscapes
    -0.63
    anni
    -0.63
    redients
    -0.62
    VERTISEMENT
    -0.62
     Stories
    -0.61
    POSITIVE LOGITS
     bother
    0.94
     hesitate
    0.93
     condone
    0.93
     tolerate
    0.90
     underestimate
    0.85
     pretend
    0.83
     endorse
    0.80
     regret
    0.80
     quarrel
    0.79
     anymore
    0.78
    Activations Density 0.069%

    No Known Activations