© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 26401
    Prev
    Next
    INDEX
    Explanations

    mentions of the word "embed" or variations of it in various contexts

    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
    111
    +0.13
    0.7%
    203
    +0.13
    0.7%
    0
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    386
    +0.13
    0.02
    203
    +0.13
    0.03
    221
    +0.12
    0.02
    Negative Logits
    undry
    -1.79
    hip
    -1.74
    fully
    -1.48
     secret
    -1.46
    php
    -1.45
     investigators
    -1.44
     principal
    -1.43
    htra
    -1.42
    ureus
    -1.39
     fame
    -1.39
    POSITIVE LOGITS
    assador
    1.45
    assadors
    1.45
    arta
    1.39
     Infantry
    1.38
    uesta
    1.37
    strom
    1.36
     anne
    1.35
    ©
    1.34
    weight
    1.32
    ry
    1.30
    Activations Density 0.195%

    No Known Activations