© 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. 31313
    Prev
    Next
    INDEX
    Explanations

    numeric identifiers associated with chemical compounds and their properties

    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
    74
    +0.12
    0.7%
    79
    +0.12
    0.7%
    110
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    167
    +0.12
    0.02
    374
    +0.12
    0.01
    185
    +0.11
    0.01
    Negative Logits
     cursor
    -1.35
     cart
    -1.34
     respectively
    -1.32
     dialog
    -1.30
     server
    -1.30
     direction
    -1.30
     bidding
    -1.30
    irected
    -1.28
    largest
    -1.28
     wire
    -1.27
    POSITIVE LOGITS
    ¢
    2.73
    Ĥ
    2.69
    »
    2.65
    ĥ
    2.64
    ·
    2.60
    ¸
    2.52
    ij
    2.49
    ³
    2.48
    §
    2.47
    ¯
    2.44
    Activations Density 0.080%

    No Known Activations