© 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. 14116
    Prev
    Next
    INDEX
    Explanations

    key terms and elements related to events, locations, and characters 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/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
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    431
    +0.15
    0.8%
    156
    +0.12
    0.7%
    277
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    145
    +0.15
    0.09
    431
    +0.12
    0.05
    89
    +0.12
    0.05
    Negative Logits
    ĥ½
    -2.03
    ²
    -2.01
    ço
    -1.72
    utter
    -1.70
    iella
    -1.69
    )](
    -1.68
    Ĥ¬
    -1.63
    IJ
    -1.61
    matically
    -1.60
    ĥ
    -1.58
    POSITIVE LOGITS
     Oval
    1.52
     quarters
    1.47
     Plate
    1.44
     decorations
    1.41
     creations
    1.40
     clocks
    1.40
     diary
    1.39
     profile
    1.35
     beard
    1.35
     life
    1.34
    Activations Density 1.179%

    No Known Activations