© 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. 13988
    Prev
    Next
    INDEX
    Explanations

    instances of the verb "is" and variations related to being or existence

    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
    23
    +0.24
    1.4%
    419
    +0.15
    0.9%
    304
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    23
    +0.24
    0.11
    304
    +0.15
    0.09
    336
    +0.13
    0.08
    Negative Logits
    Ŀ
    -4.46
    ±
    -4.37
    Ļª
    -4.32
    ĻĤ
    -4.17
    ł
    -4.01
    ·¸
    -3.94
    »¿
    -3.88
    ¡
    -3.85
    Ĥ¬
    -3.80
    İ
    -3.79
    POSITIVE LOGITS
     imperative
    1.90
    rael
    1.89
     advis
    1.85
     intended
    1.63
     mentioned
    1.60
     incumbent
    1.57
    olate
    1.56
    elf
    1.55
     supposed
    1.53
    abella
    1.53
    Activations Density 0.639%

    No Known Activations