© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 8-RES-JB
    6. 8066
    Prev
    Next
    INDEX
    Explanations

    sentences where the verb "is" is followed by a word that indicates a strong emotional or impactful reaction

    oai_token-act-pair · gpt-3.5-turbo

    statements of being or existence in descriptive contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     peg
    -0.77
     stabilization
    -0.72
     Awakens
    -0.70
     correction
    -0.70
    inav
    -0.68
     Came
    -0.68
     bumps
    -0.67
     Ago
    -0.64
    âĦ¢:
    -0.63
    lex
    -0.62
    POSITIVE LOGITS
     commonplace
    1.04
     punishable
    0.97
     indistinguishable
    0.91
     considered
    0.89
     rife
    0.88
     forbidden
    0.88
     reminiscent
    0.88
     deemed
    0.86
     prohibited
    0.84
     taboo
    0.84
    Activations Density 0.349%

    No Known Activations