© 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. 2-RES-SM
    11. 10610
    Prev
    Next
    INDEX
    Explanations

    specific keywords related to financial and corporate issues

    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
    271
    +0.21
    1.2%
    235
    +0.20
    1.2%
    98
    +0.17
    1.0%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    235
    +0.21
    0.14
    271
    +0.20
    -0.02
    345
    +0.17
    0.10
    Negative Logits
    ĥ½
    -2.36
    wagen
    -2.15
    »¿
    -1.67
    Īĺ
    -1.63
    ĸ´
    -1.56
     lieu
    -1.53
     basis
    -1.53
    Ļ
    -1.50
     purse
    -1.49
    ¥
    -1.48
    POSITIVE LOGITS
     himself
    1.72
    ously
    1.70
     herself
    1.48
     yourself
    1.43
    friends
    1.39
    mathrsfs
    1.36
    ὸ
    1.32
    ieve
    1.30
    àº
    1.30
    ÏģÏĮ
    1.30
    Activations Density 4.042%

    No Known Activations