© 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. 26262
    Prev
    Next
    INDEX
    Explanations

    instances of relational phrases and arguments in programming or logical statements

    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
    156
    +0.30
    1.7%
    17
    +0.23
    1.3%
    111
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    478
    +0.30
    0.00
    98
    +0.23
    -0.01
    282
    +0.13
    0.03
    Negative Logits
     lasting
    -1.61
    stem
    -1.36
     imprint
    -1.35
     DISCLAIM
    -1.33
     poured
    -1.28
    ↵
    -1.28
     ↵
    -1.28
    č↵
    -1.28
    ↵
    -1.28
    ↵
    -1.28
    POSITIVE LOGITS
    ses
    1.72
    sible
    1.54
    obia
    1.54
     <%=
    1.47
     Java
    1.40
     hypertension
    1.38
    cdot
    1.36
    stylesheet
    1.34
    igion
    1.32
    *^
    1.30
    Activations Density 0.551%

    No Known Activations