© 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. 1-RES-SM
    11. 24087
    Prev
    Next
    INDEX
    Explanations

    possessive pronouns indicating personal connection or ownership

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    223
    +0.13
    0.7%
    450
    +0.13
    0.7%
    416
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    338
    +0.13
    0.01
    454
    +0.13
    0.01
    96
    +0.11
    0.01
    Negative Logits
    adier
    -1.64
    rå
    -1.59
    argin
    -1.56
    rique
    -1.56
    aying
    -1.55
    ique
    -1.54
    ^](#
    -1.50
    rade
    -1.47
    prising
    -1.44
    ack
    -1.42
    POSITIVE LOGITS
     subsequ
    1.46
     purification
    1.46
     pleth
    1.43
     ago
    1.42
     evolution
    1.41
     capacity
    1.40
     scheme
    1.34
     bid
    1.33
     fidelity
    1.33
     heritage
    1.31
    Activations Density 0.003%

    No Known Activations