© 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. 31233
    Prev
    Next
    INDEX
    Explanations

    negative words or phrases related to self-perception and capability

    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
    419
    +0.14
    0.8%
    326
    +0.12
    0.7%
    218
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    249
    +0.14
    0.03
    495
    +0.12
    0.03
    223
    +0.10
    0.02
    Negative Logits
    iNdEx
    -1.76
    omb
    -1.73
     href
    -1.68
    astro
    -1.62
    nen
    -1.52
    mlin
    -1.51
    ikk
    -1.48
    iga
    -1.46
    forth
    -1.45
    sole
    -1.45
    POSITIVE LOGITS
     conduct
    1.51
     Conduct
    1.47
     equ
    1.42
     incub
    1.37
     recording
    1.37
    conduct
    1.36
    worth
    1.35
     induction
    1.34
     oneself
    1.32
     ascribed
    1.32
    Activations Density 0.117%

    No Known Activations