© 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
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 2-RES-SM
    11. 10810
    Prev
    Next
    INDEX
    Explanations

    educational strategies and concepts related to teaching mathematics

    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
    326
    +0.15
    0.9%
    12
    +0.15
    0.8%
    419
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.15
    0.06
    444
    +0.15
    -0.01
    489
    +0.14
    0.03
    Negative Logits
    fax
    -1.48
    located
    -1.40
    adows
    -1.37
     stably
    -1.32
    haul
    -1.32
    quito
    -1.30
    itat
    -1.28
    illas
    -1.27
    ichia
    -1.27
     hereinafter
    -1.25
    POSITIVE LOGITS
    eness
    1.51
     states
    1.43
    otyping
    1.42
     nation
    1.40
    ±
    1.37
     teacher
    1.36
    ħ
    1.34
    Ŀ
    1.33
     proud
    1.33
     struggle
    1.32
    Activations Density 5.632%

    No Known Activations