© 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. 12340
    Prev
    Next
    INDEX
    Explanations

    phrases that emphasize speed or efficiency

    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
    362
    +0.16
    0.9%
    255
    +0.15
    0.9%
    173
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    255
    +0.16
    0.03
    148
    +0.15
    0.01
    339
    +0.14
    0.02
    Negative Logits
    Ļª
    -2.66
    ĭ
    -2.62
    ·¸
    -2.52
    ĥ
    -2.52
    °
    -2.49
    ļ
    -2.46
    Ĭ
    -2.44
    ı
    -2.42
    ¦
    -2.42
    Į
    -2.38
    POSITIVE LOGITS
    ened
    2.03
    est
    1.92
    shots
    1.82
    fix
    1.64
    ening
    1.60
    cards
    1.57
    Tc
    1.54
    ely
    1.50
    ty
    1.49
    ingly
    1.47
    Activations Density 0.168%

    No Known Activations