© 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. 31928
    Prev
    Next
    INDEX
    Explanations

    references to judicial institutions and legal terminology

    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
    434
    +0.18
    1.0%
    435
    +0.15
    0.9%
    365
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    434
    +0.18
    0.05
    365
    +0.15
    0.04
    104
    +0.13
    0.03
    Negative Logits
    apine
    -1.84
     consumption
    -1.79
    leep
    -1.76
    ```
    -1.69
    amate
    -1.61
    esium
    -1.52
     young
    -1.52
    phrine
    -1.50
    mos
    -1.50
     seedlings
    -1.45
    POSITIVE LOGITS
    esy
    2.78
    piece
    2.00
    room
    1.97
    ship
    1.96
    rooms
    1.89
    member
    1.82
    ī
    1.80
    ĭ
    1.80
    Ļ
    1.77
    ©
    1.75
    Activations Density 0.122%

    No Known Activations