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 49152
    5. 8-RES_FS49152-JB
    6. 37193
    Prev
    Next
    INDEX
    Explanations

    specific organizations or brands relevant to political or social contexts

    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
    +0.17
    0.8%
    420
    +0.11
    0.5%
    461
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    461
    +0.17
    0.06
    401
    +0.11
    0.06
    563
    +0.11
    0.05
    Negative Logits
     dstg
    -0.70
    ãĥķãĤ¡
    -0.64
    ::::::::
    -0.61
    ĺħ
    -0.60
    ©¶æ¥µ
    -0.59
    NetMessage
    -0.58
    ],"
    -0.57
    frac
    -0.57
    iddy
    -0.56
    naissance
    -0.56
    POSITIVE LOGITS
     also
    1.17
     furthermore
    1.09
     therefore
    1.09
     additionally
    1.04
     moreover
    0.97
     emphasizes
    0.94
     nonetheless
    0.88
     however
    0.86
     retains
    0.86
     consists
    0.86
    Activations Density 0.287%

    No Known Activations