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. 12145
    Prev
    Next
    INDEX
    Explanations

    phrases related to public discourse and community issues

    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
    +0.13
    0.6%
    426
    +0.11
    0.5%
    333
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    426
    +0.13
    0.05
    333
    +0.11
    0.04
    430
    +0.10
    0.05
    Negative Logits
    etheless
    -0.91
     tremend
    -0.63
    ãĥij
    -0.61
     stret
    -0.60
    £ı
    -0.59
     spons
    -0.57
     helicop
    -0.56
     misunder
    -0.55
    Widget
    -0.55
    eco
    -0.55
    POSITIVE LOGITS
     said
    1.23
    said
    1.20
     says
    1.07
     wrote
    1.04
     he
    1.03
     reads
    0.98
     she
    0.95
     joked
    0.91
     writes
    0.90
     replied
    0.90
    Activations Density 0.086%

    No Known Activations