© 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. 23997
    Prev
    Next
    INDEX
    Explanations

    the infinitive form of verbs (to + verb)

    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
    23
    +0.20
    1.2%
    478
    +0.15
    0.8%
    160
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    160
    +0.20
    0.05
    186
    +0.15
    0.02
    310
    +0.14
    0.04
    Negative Logits
    illin
    -1.94
     briefs
    -1.67
    naire
    -1.63
    #:
    -1.61
    creen
    -1.61
     alone
    -1.58
    imeter
    -1.58
    illac
    -1.50
    á¾
    -1.49
    à¸ĩ
    -1.48
    POSITIVE LOGITS
    bey
    1.64
     rescue
    1.63
     return
    1.56
     sudden
    1.55
     recover
    1.49
     be
    1.48
     homogeneity
    1.47
     rid
    1.45
     recovery
    1.45
     reass
    1.44
    Activations Density 0.231%

    No Known Activations