© 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. 13906
    Prev
    Next
    INDEX
    Explanations

    the word "behind" and its context or related themes

    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
    16
    +0.14
    0.8%
    115
    +0.12
    0.7%
    443
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    16
    +0.14
    0.03
    485
    +0.12
    0.02
    119
    +0.12
    0.02
    Negative Logits
    «
    -2.08
    ester
    -1.67
    Ĺ
    -1.60
    ģ
    -1.56
    α
    -1.45
    Ķ
    -1.43
    umab
    -1.40
    §
    -1.38
     år
    -1.36
    AD
    -1.34
    POSITIVE LOGITS
    ships
    2.10
    enstein
    1.97
    sight
    1.61
    neath
    1.55
    drawn
    1.54
    stood
    1.49
    olin
    1.48
    ttes
    1.46
     scenes
    1.45
     obstacles
    1.44
    Activations Density 0.200%

    No Known Activations