© 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. 12767
    Prev
    Next
    INDEX
    Explanations

    references to database operations and structure

    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
    369
    +0.17
    0.9%
    478
    +0.12
    0.7%
    441
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    185
    +0.17
    0.09
    450
    +0.12
    0.10
    43
    +0.11
    0.09
    Negative Logits
    .*]{},
    -1.75
    ests
    -1.73
     [@
    -1.70
    *]{},
    -1.59
    eric
    -1.58
    .*]{}
    -1.53
    *]{}.
    -1.52
    ennial
    -1.47
    .[@
    -1.45
    .“
    -1.44
    POSITIVE LOGITS
    ·¸
    2.86
    ¿
    2.85
    ²
    2.46
    ¦
    2.38
    ŀ
    2.32
    ´
    2.28
    °
    2.18
    ¾
    2.18
    Ń
    2.13
    ¨
    2.05
    Activations Density 0.511%

    No Known Activations