© 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. 4-RES-SM
    11. 8865
    Prev
    Next
    INDEX
    Explanations

    emotional challenges and hardships in various contexts

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    156
    +0.16
    0.9%
    137
    +0.15
    0.9%
    232
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    137
    +0.16
    0.12
    198
    +0.15
    0.11
    359
    +0.13
    0.10
    Negative Logits
    smen
    -1.68
    avir
    -1.62
    ivers
    -1.60
    eks
    -1.57
    azole
    -1.56
     therefrom
    -1.56
    etine
    -1.54
    frey
    -1.52
     wisely
    -1.50
    esters
    -1.49
    POSITIVE LOGITS
     situation
    2.50
     thing
    2.47
     circumstances
    2.15
     things
    2.07
     affair
    2.01
     situations
    2.00
     circumstance
    1.97
     times
    1.96
     stuff
    1.96
    ly
    1.91
    Activations Density 1.117%

    No Known Activations