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 3072
    5. 8-RES_FS3072-JB
    6. 1674
    Prev
    Next
    INDEX
    Explanations

    words related to verifying or confirming information

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    725
    +0.10
    0.5%
    642
    +0.09
    0.4%
    455
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    642
    +0.10
    0.04
    3
    +0.09
    0.04
    455
    +0.09
    0.05
    Negative Logits
    onew
    -0.79
    enium
    -0.75
    psc
    -0.68
    bler
    -0.66
    theme
    -0.66
    neys
    -0.64
    cific
    -0.63
    olit
    -0.61
    anie
    -0.61
    ð
    -0.60
    POSITIVE LOGITS
     authenticity
    0.94
     suspicions
    0.94
     receipt
    0.92
    ations
    0.84
    orship
    0.83
     confirmation
    0.79
     sighting
    0.78
    atively
    0.78
     validity
    0.78
     rumours
    0.76
    Activations Density 0.616%

    No Known Activations