© 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. 31418
    Prev
    Next
    INDEX
    Explanations

    phrases indicating evaluations or opinions

    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
    408
    +0.12
    0.7%
    421
    +0.11
    0.6%
    320
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    331
    +0.12
    0.02
    300
    +0.11
    0.02
    462
    +0.10
    0.01
    Negative Logits
    Ľ
    -2.28
    »
    -2.10
    ½
    -2.09
    ª
    -2.08
    IJ
    -2.02
    ·
    -1.98
    Ĵ
    -1.96
    ĥ
    -1.93
    ±
    -1.93
    Ŀ
    -1.92
    POSITIVE LOGITS
     guilty
    1.86
    uke
    1.76
     favorably
    1.58
    abol
    1.54
     whether
    1.43
    idine
    1.42
     wrongly
    1.41
     worth
    1.41
     safe
    1.41
     fix
    1.39
    Activations Density 0.121%

    No Known Activations