© 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. 12496
    Prev
    Next
    INDEX
    Explanations

    the name "De"

    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
    295
    +0.14
    0.7%
    31
    +0.12
    0.7%
    271
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    295
    +0.14
    0.03
    31
    +0.12
    0.03
    110
    +0.12
    0.03
    Negative Logits
    s
    -1.64
    tons
    -1.61
    IENT
    -1.58
     insensitive
    -1.54
    ky
    -1.53
    dn
    -1.47
     these
    -1.43
    ny
    -1.42
    rors
    -1.39
    aurus
    -1.39
    POSITIVE LOGITS
    wey
    2.30
    feat
    2.27
    bbie
    2.13
    letion
    2.07
    bris
    2.03
    cember
    2.02
    legate
    2.00
    arest
    2.00
    ployment
    1.98
    composition
    1.95
    Activations Density 0.045%

    No Known Activations