© 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. 26266
    Prev
    Next
    INDEX
    Explanations

    declarations of constants in code

    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
    391
    +0.12
    0.7%
    20
    +0.11
    0.6%
    101
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    391
    +0.12
    0.03
    180
    +0.11
    0.03
    20
    +0.11
    0.04
    Negative Logits
     site
    -1.78
    ster
    -1.56
    work
    -1.54
    cology
    -1.53
    sters
    -1.52
    wards
    -1.51
    bond
    -1.51
    zone
    -1.50
    back
    -1.48
    ymoon
    -1.48
    POSITIVE LOGITS
    raints
    2.10
    ĻĤ
    2.10
    ellation
    2.07
    rains
    2.03
     glimps
    1.69
    Ĵ
    1.66
    ¦
    1.61
    ructions
    1.56
    ricted
    1.53
    expr
    1.53
    Activations Density 0.118%

    No Known Activations