© 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. 30736
    Prev
    Next
    INDEX
    Explanations

    phrases or words emphasizing the concept of "never" or the idea of negation

    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
    358
    +0.13
    0.7%
    476
    +0.11
    0.6%
    492
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    206
    +0.13
    0.03
    132
    +0.11
    0.03
    118
    +0.11
    0.04
    Negative Logits
    Ħ
    -2.03
    ¦
    -2.01
    ©
    -1.79
    ³
    -1.79
    opolys
    -1.67
    ¯
    -1.57
     each
    -1.55
    ization
    -1.54
    ¶
    -1.53
    ķ
    -1.52
    POSITIVE LOGITS
    theless
    2.78
     regret
    1.69
    yy
    1.68
    hit
    1.63
    married
    1.61
    ogs
    1.59
     rocket
    1.58
     remar
    1.53
    kill
    1.50
    MATCH
    1.49
    Activations Density 0.121%

    No Known Activations