© 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. 11615
    Prev
    Next
    INDEX
    Explanations

    references to abortion

    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
    410
    +0.16
    0.9%
    21
    +0.14
    0.8%
    78
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    410
    +0.16
    0.03
    443
    +0.14
    0.03
    241
    +0.13
    0.03
    Negative Logits
    fully
    -2.18
     newsletters
    -1.57
    hips
    -1.52
    izations
    -1.47
     terms
    -1.46
    izable
    -1.46
    chle
    -1.40
    pieces
    -1.38
     parts
    -1.36
    ESULT
    -1.35
    POSITIVE LOGITS
    oard
    2.50
    yrinth
    2.49
    brev
    2.19
    ytes
    2.17
    idopsis
    2.13
    by
    2.09
    dominal
    2.01
    ulum
    1.94
    urd
    1.93
    uchi
    1.92
    Activations Density 0.107%

    No Known Activations