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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 4-RES-SM
    6. 11848
    Prev
    Next
    INDEX
    Explanations

    quantities and large numbers related to populations and metrics

    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
    348
    +0.20
    1.3%
    466
    +0.13
    0.8%
    78
    +0.13
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    348
    +0.20
    0.04
    502
    +0.13
    0.04
    466
    +0.13
    0.04
    Negative Logits
     direction
    -1.63
    vist
    -1.59
     modo
    -1.54
    schemas
    -1.45
     slightest
    -1.41
    anium
    -1.38
    lasses
    -1.35
     vice
    -1.33
    aes
    -1.28
    rail
    -1.26
    POSITIVE LOGITS
     digit
    2.16
     of
    2.01
    »¿
    1.59
     owed
    1.54
    lotte
    1.50
    Äĩ
    1.49
    ĭ
    1.49
     cheering
    1.47
    imeter
    1.47
     upon
    1.47
    Activations Density 0.243%

    No Known Activations