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. 12186
    Prev
    Next
    INDEX
    Explanations

    prepositions and phrases related to conditions and contexts

    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
    398
    +0.10
    0.5%
    608
    +0.10
    0.5%
    511
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    608
    +0.10
    0.01
    281
    +0.10
    0.01
    511
    +0.09
    0.01
    Negative Logits
    irez
    -0.81
    å§
    -0.80
    apest
    -0.79
    sis
    -0.78
    soever
    -0.76
     largeDownload
    -0.76
    TAIN
    -0.75
    ãĥ¢
    -0.75
    GREEN
    -0.73
    ģĸ
    -0.73
    POSITIVE LOGITS
     inflation
    1.15
     biases
    1.05
     inconsistencies
    1.04
     differences
    1.03
     uncertainties
    0.98
     factors
    0.98
     variables
    0.96
     confounding
    0.96
     discrepancies
    0.95
     weaknesses
    0.93
    Activations Density 0.030%

    No Known Activations