© 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. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 3-RES-JB
    11. 17461
    Prev
    Next
    INDEX
    Explanations

    phrases related to following or adhering to rules and instructions

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "according" and variations related to reasoning or conditional statements

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Demons
    -0.63
     Osc
    -0.61
    apons
    -0.61
     Sahara
    -0.61
    eg
    -0.60
    tan
    -0.60
    blast
    -0.60
    arks
    -0.59
    oldemort
    -0.58
    auri
    -0.58
    POSITIVE LOGITS
    chwitz
    0.79
    tains
    0.78
    ety
    0.75
    format
    0.74
    Style
    0.71
    graded
    0.68
    itiz
    0.67
     corresponds
    0.66
    tained
    0.65
    achev
    0.65
    Activations Density 0.044%

    No Known Activations