© 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 12288
    9. 8-RES_FS12288-JB
    11. 9760
    Prev
    Next
    INDEX
    Explanations

    phrases related to investigations and probes

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    549
    +0.13
    0.6%
    447
    +0.12
    0.6%
    113
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    744
    +0.13
    0.04
    549
    +0.12
    0.04
    113
    +0.11
    0.03
    Negative Logits
    emonic
    -0.72
    licks
    -0.67
    birth
    -0.67
    aez
    -0.62
     goodbye
    -0.61
     maiden
    -0.60
     penn
    -0.60
    yip
    -0.57
    ãĤª
    -0.57
    ane
    -0.57
    POSITIVE LOGITS
     probing
    1.04
     investigation
    0.91
     into
    0.91
     involving
    0.90
     undertaken
    0.88
     inquiry
    0.88
     investigating
    0.87
     conducted
    0.85
     probe
    0.85
     Probe
    0.85
    Activations Density 0.107%

    No Known Activations