© 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. 11990
    Prev
    Next
    INDEX
    Explanations

    negative descriptors related to cultural or ethnic reference points

    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
    21
    +0.23
    1.3%
    450
    +0.16
    0.9%
    203
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    450
    +0.23
    0.04
    203
    +0.16
    0.05
    313
    +0.14
    0.05
    Negative Logits
     LICENSE
    -1.61
    ctions
    -1.61
     silence
    -1.56
     diagrams
    -1.45
     histograms
    -1.45
    herty
    -1.45
     electrophoresis
    -1.43
    cket
    -1.42
     kinetics
    -1.39
    LLY
    -1.38
    POSITIVE LOGITS
    esque
    2.41
    dominated
    2.12
    sponsored
    2.07
    occupied
    1.93
    supported
    1.91
    inspired
    1.88
    owned
    1.87
    speaking
    1.87
    approved
    1.86
    based
    1.85
    Activations Density 0.196%

    No Known Activations