© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 2-RES-SM
    11. 31811
    Prev
    Next
    INDEX
    Explanations

    conjunctions and expressions of agreement or addition

    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/2-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.2.hook_resid_post
    Hook Layer
    2
    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
    50
    +0.15
    0.8%
    446
    +0.12
    0.7%
    436
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    50
    +0.15
    0.08
    494
    +0.12
    0.06
    15
    +0.12
    0.08
    Negative Logits
    ιÏĤ
    -1.91
    ี
    -1.63
    éĩĮ
    -1.61
    clud
    -1.56
    fecture
    -1.54
    ение
    -1.54
    ylum
    -1.54
    oder
    -1.52
    was
    -1.50
    emia
    -1.43
    POSITIVE LOGITS
    rogens
    2.10
    naments
    1.83
     executed
    1.57
    UX
    1.54
    che
    1.51
    rog
    1.47
    esp
    1.45
    ech
    1.45
     otherwise
    1.38
     rue
    1.36
    Activations Density 0.588%

    No Known Activations