© 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. 4-RES-SM
    11. 9490
    Prev
    Next
    INDEX
    Explanations

    references to user interactions and comments in a discussion context

    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
    221
    +0.12
    0.7%
    365
    +0.12
    0.6%
    331
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    365
    +0.12
    0.04
    89
    +0.12
    0.02
    509
    +0.11
    0.04
    Negative Logits
    ķ
    -1.99
    ¸
    -1.91
    ¹
    -1.83
    Ħ
    -1.73
    ĭ
    -1.69
    ı
    -1.69
    ļ
    -1.68
    ·
    -1.68
    ³
    -1.64
    ľ
    -1.55
    POSITIVE LOGITS
    aries
    2.38
    naire
    2.29
    ariat
    2.13
    ators
    2.08
     comment
    1.95
    ership
    1.94
     moderation
    1.93
    arial
    1.93
    ative
    1.91
    feedback
    1.88
    Activations Density 0.386%

    No Known Activations