© 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. 14413
    Prev
    Next
    INDEX
    Explanations

    references to animals, particularly those related to bovine and other livestock

    oai_token-act-pair · gpt-4o-mini
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    156
    +0.22
    1.2%
    34
    +0.13
    0.7%
    446
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    446
    +0.22
    0.01
    304
    +0.13
    0.01
    373
    +0.12
    0.01
    Negative Logits
    vised
    -1.50
     direction
    -1.45
     geared
    -1.44
     starred
    -1.38
     variance
    -1.38
    ributed
    -1.37
     jury
    -1.36
    uated
    -1.36
     posted
    -1.33
    Seq
    -1.31
    POSITIVE LOGITS
    ĻĤ
    3.58
    Ļª
    3.18
    ¿½
    3.14
    ¨
    3.08
    «
    3.07
    ł
    3.06
    Ģ
    3.03
    »¿
    2.99
    ĸ´
    2.99
    ¬
    2.97
    Activations Density 0.074%

    No Known Activations