© 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. 4-RES-SM
    11. 9695
    Prev
    Next
    INDEX
    Explanations

    geographical locations and events within Colorado

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    307
    +0.14
    0.8%
    137
    +0.13
    0.7%
    75
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    216
    +0.14
    0.07
    60
    +0.13
    0.02
    145
    +0.12
    0.03
    Negative Logits
     tumour
    -2.58
     £
    -2.49
     programme
    -2.36
     harbour
    -2.34
     programmes
    -2.27
     tumours
    -2.26
     fibres
    -2.25
     Liverpool
    -2.11
     fibre
    -2.09
     realised
    -2.05
    POSITIVE LOGITS
     Colorado
    4.26
     Nebraska
    4.14
     Iowa
    4.06
     Utah
    3.90
     Denver
    3.88
     Wyoming
    3.87
     Montana
    3.77
     Missouri
    3.72
     Kansas
    3.70
     Idaho
    3.68
    Activations Density 1.092%

    No Known Activations