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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 24576
    5. 8-RES_FS24576-JB
    6. 8633
    Prev
    Next
    INDEX
    Explanations

    dialogue punctuated by exclamation points and quotes

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    373
    +0.18
    0.8%
    481
    +0.12
    0.5%
    678
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    495
    +0.18
    0.04
    222
    +0.12
    0.03
    668
    +0.10
    0.03
    Negative Logits
    artifacts
    -1.01
    etheless
    -0.82
    theless
    -0.82
    icultural
    -0.78
    rats
    -0.78
    icter
    -0.77
    creation
    -0.76
    hesda
    -0.76
    clusions
    -0.74
    earable
    -0.73
    POSITIVE LOGITS
     sounding
    0.92
     sarcast
    0.91
     aloud
    0.90
     loudly
    0.90
     loud
    0.88
     louder
    0.86
     softly
    0.85
     yelled
    0.83
     whispered
    0.82
     smir
    0.81
    Activations Density 0.147%

    No Known Activations