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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 12288
    5. 8-RES_FS12288-JB
    6. 180
    Prev
    Next
    INDEX
    Explanations

    declarative statements expressing certainty

    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
    373
    +0.14
    0.6%
    372
    +0.10
    0.5%
    70
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    8
    +0.14
    0.04
    750
    +0.10
    0.04
    70
    +0.10
    0.04
    Negative Logits
    isco
    -0.81
    mage
    -0.75
    aukee
    -0.73
    gencies
    -0.69
    issance
    -0.69
    earable
    -0.68
    aunder
    -0.66
    acies
    -0.65
    pload
    -0.65
     consum
    -0.64
    POSITIVE LOGITS
     firsthand
    0.99
     anecd
    0.82
     plenty
    0.82
     how
    0.77
     why
    0.76
     nothing
    0.74
     personally
    0.71
     what
    0.69
     lots
    0.68
     exactly
    0.65
    Activations Density 0.104%

    No Known Activations