© 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 EvalsExportsAPI 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. 30461
    Prev
    Next
    INDEX
    Explanations

    instances of the word "writing."

    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
    74
    +0.16
    0.9%
    230
    +0.10
    0.6%
    71
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    74
    +0.16
    0.02
    175
    +0.10
    0.02
    222
    +0.10
    0.02
    Negative Logits
    ľĵ
    -1.92
    odor
    -1.72
    ®
    -1.61
    ]));
    -1.47
    lights
    -1.45
    '))
    -1.44
    Ĺ
    -1.43
    ands
    -1.42
    >>>
    -1.40
    iona
    -1.36
    POSITIVE LOGITS
     sentences
    1.41
    achus
    1.39
     waiver
    1.37
    ktop
    1.35
    ware
    1.33
     unpublished
    1.32
     sentence
    1.30
     rule
    1.27
     Symphony
    1.27
    fellow
    1.24
    Activations Density 0.047%

    No Known Activations