© 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. 14324
    Prev
    Next
    INDEX
    Explanations

    mentions of family relationships and themes related to familial roles and dynamics

    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
    72
    +0.17
    1.0%
    34
    +0.14
    0.8%
    320
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    34
    +0.17
    0.14
    72
    +0.14
    0.07
    411
    +0.13
    0.06
    Negative Logits
    alth
    -1.87
    ãĤĩ
    -1.74
    imb
    -1.70
    orems
    -1.62
    常
    -1.51
    tras
    -1.47
    ubot
    -1.44
    áĢº
    -1.43
    roidism
    -1.42
    ãĤį
    -1.41
    POSITIVE LOGITS
     Cup
    1.42
     vote
    1.40
     voting
    1.36
     VIP
    1.33
     decides
    1.30
     coffin
    1.27
     announces
    1.27
     announce
    1.27
    mate
    1.26
     Augustus
    1.26
    Activations Density 2.947%

    No Known Activations