© 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. 12266
    Prev
    Next
    INDEX
    Explanations

    commands and settings related to configuration changes

    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
    423
    +0.18
    1.1%
    17
    +0.17
    1.0%
    419
    +0.16
    1.0%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    17
    +0.18
    0.01
    423
    +0.17
    0.07
    419
    +0.16
    0.04
    Negative Logits
    ĻĤ
    -4.49
    ²
    -3.17
    ¤
    -3.09
    ª
    -3.07
    ¼
    -3.01
    ¿½
    -2.98
    ı
    -2.96
    İ
    -2.93
    Ī
    -2.89
    ł
    -2.86
    POSITIVE LOGITS
     parameters
    1.71
     deadlines
    1.67
     alarms
    1.65
     basis
    1.62
     alarm
    1.60
     accordingly
    1.54
     schedule
    1.50
     variables
    1.48
     mates
    1.46
     priorities
    1.45
    Activations Density 0.831%

    No Known Activations