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

    Neuronpedia

    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. 4-RES-SM
    11. 12180
    Prev
    Next
    INDEX
    Explanations

    user interface elements related to selection and dropdown menus

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    307
    +0.14
    0.8%
    495
    +0.11
    0.6%
    17
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    17
    +0.14
    0.01
    192
    +0.11
    0.00
    277
    +0.10
    0.01
    Negative Logits
    usalem
    -1.64
    poons
    -1.57
    ĻĤ
    -1.57
    ness
    -1.42
    pes
    -1.41
    loo
    -1.34
    oux
    -1.32
    nier
    -1.29
    gie
    -1.28
    geois
    -1.27
    POSITIVE LOGITS
    ors
    2.39
    ively
    2.32
    able
    2.23
    iveness
    2.12
    ivity
    2.11
    ableView
    1.88
    rons
    1.82
    ive
    1.80
    urable
    1.80
    ivities
    1.79
    Activations Density 0.090%

    No Known Activations