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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 768
    9. 8-RES_FS768-JB
    11. 652
    Prev
    Next
    INDEX
    Explanations

    phrases expressing gratitude and appreciation

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    259
    +0.11
    0.5%
    381
    +0.10
    0.5%
    63
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.11
    0.20
    424
    +0.10
    0.09
    317
    +0.10
    0.11
    Negative Logits
    aeda
    -0.67
    ulhu
    -0.65
    rooms
    -0.65
    apse
    -0.65
    iche
    -0.64
    arettes
    -0.64
    acid
    -0.64
     Lans
    -0.63
    oto
    -0.63
    scape
    -0.62
    POSITIVE LOGITS
    giving
    1.46
    FUL
    0.96
     acknowled
    0.95
    eous
    0.92
    ably
    0.91
    ifully
    0.88
    ledged
    0.84
     bestowed
    0.83
    fully
    0.80
     recipient
    0.79
    Activations Density 5.691%

    No Known Activations