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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & 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. 12188
    Prev
    Next
    INDEX
    Explanations

    instances of the verb "to be" in different tenses, particularly in relation to reports of injuries or casualties

    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
    418
    +0.12
    0.5%
    566
    +0.12
    0.5%
    102
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    102
    +0.12
    0.05
    566
    +0.12
    0.04
    418
    +0.11
    0.04
    Negative Logits
    inventoryQuantity
    -0.89
     Gutenberg
    -0.83
     Solution
    -0.82
     Achieve
    -0.75
     Grass
    -0.75
     Scrib
    -0.75
     Gleaming
    -0.73
     Dragonbound
    -0.72
    ciation
    -0.71
    sonian
    -0.69
    POSITIVE LOGITS
     hospitalized
    1.32
     injured
    1.31
     killed
    1.20
     arrested
    1.17
     detained
    1.16
     evacuated
    1.12
     wounded
    1.04
     murdered
    1.03
     ejected
    1.01
     interviewed
    0.98
    Activations Density 0.089%

    No Known Activations