© 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. 27507
    Prev
    Next
    INDEX
    Explanations

    references to family and personal relationships

    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
    56
    +0.15
    0.9%
    17
    +0.14
    0.8%
    183
    +0.13
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    56
    +0.15
    0.00
    183
    +0.14
    0.06
    379
    +0.13
    0.08
    Negative Logits
    ĥ
    -2.41
    ĥ½
    -2.02
    Īĺ
    -1.87
    İ
    -1.79
    ¸
    -1.71
    Ĥ¬
    -1.70
    ľĵ
    -1.67
    ½
    -1.66
    ¯
    -1.63
    ĵ
    -1.61
    POSITIVE LOGITS
     goal
    1.50
    ickets
    1.44
     goodbye
    1.43
    stic
    1.38
     ahead
    1.38
    vous
    1.37
     Cookie
    1.32
     goals
    1.31
     plans
    1.28
    buntu
    1.27
    Activations Density 0.704%

    No Known Activations