© 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 49152
    9. 8-RES_FS49152-JB
    11. 37390
    Prev
    Next
    INDEX
    Explanations

    references to specific geographical locations, particularly the word "Peninsula."

    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
    288
    +0.17
    0.8%
    99
    +0.11
    0.5%
    547
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    547
    +0.17
    0.01
    288
    +0.11
    0.01
    537
    +0.10
    0.01
    Negative Logits
    thumbnails
    -0.93
    ãĥīãĥ©
    -0.85
    nell
    -0.71
    hig
    -0.71
    ¼
    -0.71
    Interstitial
    -0.70
    Strike
    -0.69
    cham
    -0.69
    ¯
    -0.67
    Male
    -0.67
    POSITIVE LOGITS
     peninsula
    1.20
     Peninsula
    1.13
    side
    0.86
     ridge
    0.83
     region
    0.80
     Rim
    0.74
     rim
    0.72
    wide
    0.71
    itect
    0.71
     Tunisia
    0.69
    Activations Density 0.005%

    No Known Activations