© 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. 13420
    Prev
    Next
    INDEX
    Explanations

    mathematical expressions and operations involving variables and constants

    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
    392
    +0.13
    0.7%
    406
    +0.12
    0.7%
    369
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    392
    +0.13
    0.02
    147
    +0.12
    0.01
    506
    +0.12
    0.01
    Negative Logits
    ^âĪĴ
    -1.66
    soever
    -1.64
    merce
    -1.61
    ijer
    -1.55
    ^--
    -1.55
    ;}
    -1.46
    ortality
    -1.43
    usement
    -1.43
    uscular
    -1.42
    ^-
    -1.34
    POSITIVE LOGITS
    anza
    1.41
     arsenal
    1.40
     boast
    1.39
     repertoire
    1.37
     boasts
    1.35
     lies
    1.34
     lesson
    1.31
     formula
    1.30
     vocabulary
    1.30
     parl
    1.29
    Activations Density 0.047%

    No Known Activations