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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & 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. 12200
    Prev
    Next
    INDEX
    Explanations

    the presence of expressions related to gratitude and awareness of support

    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
    447
    +0.24
    1.1%
    481
    +0.20
    0.9%
    640
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.24
    0.04
    6
    +0.20
    0.04
    156
    +0.13
    0.03
    Negative Logits
     intens
    -0.68
    oreal
    -0.65
    daq
    -0.63
     culminating
    -0.62
    odied
    -0.60
    ipel
    -0.60
    orate
    -0.60
    velt
    -0.59
     aven
    -0.58
     convoy
    -0.57
    POSITIVE LOGITS
     Firstly
    1.22
     Whilst
    1.07
    Firstly
    1.02
     WHY
    0.94
     THERE
    0.94
     Regarding
    0.94
     Whenever
    0.93
     NEVER
    0.86
     Contrary
    0.84
     There
    0.84
    Activations Density 0.126%

    No Known Activations