© 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. 4-RES-SM
    11. 13895
    Prev
    Next
    INDEX
    Explanations

    phrases expressing evaluation or assessment of something's status or value

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    348
    +0.19
    1.0%
    64
    +0.12
    0.7%
    330
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    300
    +0.19
    0.02
    289
    +0.12
    0.02
    83
    +0.11
    0.04
    Negative Logits
    ±
    -2.14
    abouts
    -1.67
    UTERS
    -1.59
    Īĺ
    -1.54
    ments
    -1.53
    eon
    -1.50
    rish
    -1.47
    °
    -1.47
    zin
    -1.47
    Ħ
    -1.47
    POSITIVE LOGITS
     acceptable
    2.13
     authoritative
    2.02
     harmless
    2.00
     unsuitable
    1.91
     unsafe
    1.90
     conclusive
    1.84
     \'
    1.83
     "
    1.82
     unacceptable
    1.80
     trustworthy
    1.80
    Activations Density 0.292%

    No Known Activations