© 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. 26088
    Prev
    Next
    INDEX
    Explanations

    negative connotations or phrases related to pre-existing conditions and their impacts

    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
    419
    +0.13
    0.7%
    218
    +0.12
    0.6%
    397
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    235
    +0.13
    0.03
    50
    +0.12
    0.02
    115
    +0.12
    0.02
    Negative Logits
    ESULT
    -1.95
    inux
    -1.93
    ictionaries
    -1.74
    ILITY
    -1.69
    plist
    -1.61
    idelines
    -1.61
    cias
    -1.57
    OOGLE
    -1.57
    "></
    -1.49
    %"
    -1.49
    POSITIVE LOGITS
    Ĥ
    1.88
    IJ
    1.85
    ¿
    1.83
    ¾
    1.79
    ¬
    1.66
    )],
    1.65
     ourselves
    1.59
    Ĥ¬
    1.58
    mortem
    1.58
    Ħ
    1.57
    Activations Density 0.075%

    No Known Activations