© 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. 31096
    Prev
    Next
    INDEX
    Explanations

    references to software and programming constructs

    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
    54
    +0.11
    0.6%
    379
    +0.10
    0.6%
    460
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    379
    +0.11
    0.03
    369
    +0.10
    0.01
    467
    +0.10
    0.01
    Negative Logits
    ¿½
    -3.02
    ı
    -2.70
    Ń
    -2.56
    Ĭ
    -2.55
    ł
    -2.50
    Ħ
    -2.44
    ³
    -2.42
    ¨
    -2.35
    ´
    -2.32
    ¢
    -2.22
    POSITIVE LOGITS
    ymphony
    1.91
    ollen
    1.62
    FIL
    1.59
    idone
    1.57
    album
    1.53
    gado
    1.46
    quito
    1.44
    idopsis
    1.39
    iewicz
    1.39
    starter
    1.38
    Activations Density 0.079%

    No Known Activations