© 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. 13238
    Prev
    Next
    INDEX
    Explanations

    content related to care home services and community-oriented philosophies

    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
    326
    +0.21
    1.2%
    207
    +0.14
    0.8%
    252
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    23
    +0.21
    0.16
    326
    +0.14
    0.06
    252
    +0.13
    0.11
    Negative Logits
    endas
    -1.59
    };
    -1.55
    ño
    -1.53
    amento
    -1.50
     hour
    -1.46
     Wrote
    -1.45
    veis
    -1.40
    imiento
    -1.37
    endl
    -1.36
     drafts
    -1.36
    POSITIVE LOGITS
    ¿½
    3.05
    ĸ´
    2.87
    ´
    2.73
    ħ
    2.72
    ĥ
    2.66
    ij
    2.66
    º
    2.64
    Ī
    2.62
    Ĥ¬
    2.62
    Ń
    2.60
    Activations Density 3.971%

    No Known Activations