© 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. 0-RES-JB
    11. 13069
    Prev
    Next
    INDEX
    Explanations

    words related to impressing others

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the word "impress" and related sentiments

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Disabled
    -0.79
    GREEN
    -0.77
    called
    -0.72
    oken
    -0.67
     Ending
    -0.66
    aka
    -0.65
    Brave
    -0.65
     Term
    -0.63
     Removal
    -0.63
    held
    -0.63
    POSITIVE LOGITS
     impress
    3.66
     embarrass
    1.70
     disappoint
    1.64
     annoy
    1.48
     amuse
    1.23
     disturb
    1.18
     obsess
    1.16
     excel
    1.15
     dazz
    1.12
     fright
    1.12
    Activations Density 0.016%

    No Known Activations