© 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. 12550
    Prev
    Next
    INDEX
    Explanations

    punctuation and symbols within code or structured text

    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
    0
    +0.18
    1.0%
    451
    +0.12
    0.7%
    159
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    0
    +0.18
    0.06
    503
    +0.12
    0.04
    157
    +0.12
    0.05
    Negative Logits
    itud
    -1.71
    raf
    -1.65
    iste
    -1.63
    äsident
    -1.57
    sic
    -1.57
    icum
    -1.50
    azioni
    -1.48
    amycin
    -1.47
    ophyll
    -1.46
     lasting
    -1.44
    POSITIVE LOGITS
    xspace
    1.69
    ¦
    1.67
     tall
    1.66
    µ
    1.65
    iece
    1.58
    «
    1.52
    ´
    1.50
     suff
    1.48
    ²
    1.47
    ¢
    1.46
    Activations Density 0.288%

    No Known Activations