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 24576
    5. 8-RES_FS24576-JB
    6. 829
    Prev
    Next
    INDEX
    Explanations

    phrases related to government actions or policies

    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
    141
    +0.12
    0.5%
    102
    +0.10
    0.5%
    86
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    141
    +0.12
    0.02
    582
    +0.10
    0.02
    55
    +0.10
    0.03
    Negative Logits
     Practices
    -0.76
    anwhile
    -0.71
    anu
    -0.68
     Strategies
    -0.67
     cabinets
    -0.67
    zag
    -0.67
     Companies
    -0.67
    quist
    -0.67
    ahoo
    -0.66
    enger
    -0.66
    POSITIVE LOGITS
     represented
    0.88
     imperfect
    0.85
     lacking
    0.83
     presented
    0.80
     considered
    0.79
     contained
    0.79
     depicted
    0.78
     portrayed
    0.78
     technically
    0.77
     admittedly
    0.77
    Activations Density 0.095%

    No Known Activations