© 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 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. 10419
    Prev
    Next
    INDEX
    Explanations

    SQL commands and their structure in database creation statements

    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
    185
    +0.12
    0.6%
    415
    +0.11
    0.6%
    47
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    157
    +0.12
    0.06
    298
    +0.11
    0.05
    431
    +0.11
    0.03
    Negative Logits
     deserve
    -1.66
    advant
    -1.52
     developing
    -1.47
    photos
    -1.46
     responsible
    -1.45
     manuscript
    -1.44
    doi
    -1.42
     worthy
    -1.41
    blockList
    -1.41
    rapeut
    -1.39
    POSITIVE LOGITS
    OUBLE
    1.81
    AIN
    1.74
    ORD
    1.73
    UMP
    1.70
    ITION
    1.70
    ONG
    1.66
     OUT
    1.64
    BIN
    1.61
    ellation
    1.56
     CIRCUIT
    1.56
    Activations Density 0.376%

    No Known Activations