© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 1-RES-SM
    11. 27686
    Prev
    Next
    INDEX
    Explanations

    instances of the word "for," suggesting a focus on purpose or intended use

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/1-res-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.1.hook_resid_post
    Hook Layer
    1
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    478
    +0.19
    1.1%
    510
    +0.15
    0.8%
    434
    +0.13
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    69
    +0.19
    0.07
    150
    +0.15
    0.05
    428
    +0.13
    0.04
    Negative Logits
    irl
    -1.84
    avel
    -1.67
    mare
    -1.58
    ainer
    -1.57
    де
    -1.44
     sense
    -1.44
    andez
    -1.43
    ulp
    -1.42
     rinsed
    -1.41
     synthesized
    -1.41
    POSITIVE LOGITS
    ½
    2.46
    Ĭ
    2.39
    «
    2.37
    Ł
    2.22
    ĸ´
    2.14
    ¼
    1.97
    ª
    1.97
    ı
    1.95
    Īĺ
    1.94
     use
    1.90
    Activations Density 0.222%

    No Known Activations