© 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. 8927
    Prev
    Next
    INDEX
    Explanations

    prepositions and phrases related to location and time

    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
    307
    +0.28
    1.6%
    306
    +0.15
    0.9%
    274
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    307
    +0.28
    0.09
    139
    +0.15
    0.08
    69
    +0.14
    0.07
    Negative Logits
    edly
    -1.56
     anytime
    -1.55
    waves
    -1.48
     throughout
    -1.43
     worden
    -1.43
    heets
    -1.42
     everywhere
    -1.42
    illin
    -1.42
    rogens
    -1.42
     exceptions
    -1.38
    POSITIVE LOGITS
    ĥ½
    2.96
    µ
    2.82
     which
    2.54
    ³
    2.33
    ·¸
    2.30
    ©
    2.26
    ±
    2.20
    »
    2.15
    £
    2.13
    Ħ
    2.12
    Activations Density 0.602%

    No Known Activations