© 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
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 25824
    Prev
    Next
    INDEX
    Explanations

    negations or terms indicating denial or absence

    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
    603
    +0.10
    0.4%
    431
    +0.09
    0.4%
    373
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    696
    +0.10
    0.05
    431
    +0.09
    0.04
    212
    +0.09
    0.04
    Negative Logits
    interstitial
    -0.72
     unprotected
    -0.69
     redes
    -0.67
     Classification
    -0.66
     princ
    -0.63
     Casting
    -0.63
     Accessed
    -0.62
    ipel
    -0.62
    oided
    -0.59
     embark
    -0.58
    POSITIVE LOGITS
    't
    1.53
    ned
    1.00
    ÃŃ
    0.93
    etsk
    0.90
    ates
    0.86
    ´
    0.84
    uts
    0.83
    eness
    0.83
    nit
    0.79
    zed
    0.78
    Activations Density 0.068%

    No Known Activations