© 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. 7-RES-JB
    11. 24106
    Prev
    Next
    INDEX
    Explanations

    the verb "possess" along with related noun forms

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the word "possess" and its variations

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 7-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    7
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    reddits
    -0.70
    bank
    -0.69
    mil
    -0.68
     externalToEVAOnly
    -0.64
    MO
    -0.63
    aver
    -0.61
     laundering
    -0.61
     Dough
    -0.59
    lins
    -0.59
    edu
    -0.58
    POSITIVE LOGITS
     faculties
    0.83
    ership
    0.82
     tremend
    0.80
    iveness
    0.79
     capacities
    0.77
     possession
    0.77
    ively
    0.77
     ambitions
    0.77
     considerable
    0.76
     sufficient
    0.76
    Activations Density 0.062%

    No Known Activations