© 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. 29962
    Prev
    Next
    INDEX
    Explanations

    references to medication usage and its effects

    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/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
    Neuron Alignment
    Index
    Value
    % of L₁
    320
    +0.13
    0.7%
    368
    +0.12
    0.7%
    271
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    418
    +0.13
    0.00
    241
    +0.12
    0.03
    495
    +0.10
    0.01
    Negative Logits
     excav
    -1.80
     erected
    -1.79
     storms
    -1.70
     Township
    -1.68
     Sew
    -1.68
     monuments
    -1.67
    athered
    -1.64
     Spring
    -1.62
     formations
    -1.60
     meteor
    -1.60
    POSITIVE LOGITS
     dose
    2.20
     efficacy
    2.06
     treatment
    2.05
     drug
    2.03
     doses
    1.81
     therapy
    1.79
    Ī
    1.76
     dosage
    1.74
     option
    1.73
     prophylaxis
    1.72
    Activations Density 0.461%

    No Known Activations