© 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. 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. 4374
    Prev
    Next
    INDEX
    Explanations

    mentions of the word "blanche" in various contexts

    oai_token-act-pair · gpt-3.5-turbo

    references to a specific person, or that person's name and various contexts surrounding them

    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
     ejac
    -0.70
    APS
    -0.70
    anthrop
    -0.69
    isSpecialOrderable
    -0.69
    INST
    -0.66
    Else
    -0.65
    iating
    -0.65
    ODUCT
    -0.65
    soDeliveryDate
    -0.64
    iage
    -0.64
    POSITIVE LOGITS
    eers
    1.11
    anche
    0.95
    erness
    0.84
    ttes
    0.81
    ktop
    0.80
    irie
    0.79
    tta
    0.79
    xual
    0.78
    iflower
    0.73
    zza
    0.73
    Activations Density 0.014%

    No Known Activations