© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 11241
    Prev
    Next
    INDEX
    Explanations

    prepositions, particularly "in" and variations related to position or location

    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
    +0.35
    2.1%
    478
    +0.18
    1.1%
    274
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    166
    +0.35
    0.04
    342
    +0.18
    0.04
    135
    +0.12
    0.05
    Negative Logits
    edly
    -2.02
    ģ
    -1.79
    heid
    -1.72
    heads
    -1.67
    tailed
    -1.55
    »¿
    -1.53
    ment
    -1.52
    ij
    -1.52
    ĵ
    -1.50
    amycin
    -1.50
    POSITIVE LOGITS
     order
    2.76
     addition
    2.29
    clud
    2.24
    frastructure
    2.20
     accordance
    2.20
     terms
    2.18
    animate
    2.15
     vain
    2.15
     lieu
    2.14
     conjunction
    2.06
    Activations Density 0.217%

    No Known Activations