© 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 EvalsExportsinterp-engineNEWAPI 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. 31436
    Prev
    Next
    INDEX
    Explanations

    phrases containing the word "up," often in various contexts related to improvement or resolution

    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
    73
    +0.12
    0.7%
    347
    +0.11
    0.7%
    126
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    73
    +0.12
    0.04
    306
    +0.11
    0.04
    292
    +0.11
    0.04
    Negative Logits
    ĻĤ
    -3.88
    Īĺ
    -3.17
    ĭ
    -3.14
    ī
    -3.00
    ĸ´
    -2.87
    ĥ
    -2.82
    Ĩ
    -2.81
    ĸ
    -2.75
    ·
    -2.73
    İ
    -2.71
    POSITIVE LOGITS
    antly
    1.54
    ively
    1.50
    opically
    1.49
    istically
    1.48
    stant
    1.45
    ificantly
    1.42
    yi
    1.39
    ionic
    1.39
    entially
    1.34
    ivat
    1.31
    Activations Density 0.551%

    No Known Activations