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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 23866
    Prev
    Next
    INDEX
    Explanations

    instances of the word "return" in various forms

    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
    654
    +0.09
    0.4%
    619
    +0.09
    0.4%
    322
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    654
    +0.09
    0.03
    469
    +0.09
    0.02
    619
    +0.08
    0.03
    Negative Logits
    ussen
    -0.73
    inen
    -0.69
     disse
    -0.65
    bis
    -0.64
    ghai
    -0.63
     gauge
    -0.60
    asm
    -0.59
    Cola
    -0.58
    bian
    -0.58
    chen
    -0.58
    POSITIVE LOGITS
     home
    0.99
    home
    0.89
     HOME
    0.88
    ees
    0.87
     safely
    0.75
    ported
    0.70
    nil
    0.68
     postage
    0.68
     Home
    0.67
    gently
    0.67
    Activations Density 0.025%

    No Known Activations