© 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. 12871
    Prev
    Next
    INDEX
    Explanations

    words related to the conclusion or final part of something, often indicated by the term "Finale."

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "Fin."

    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
     rodents
    -0.72
     zoo
    -0.71
    YY
    -0.70
    ounding
    -0.67
     privileges
    -0.66
    amed
    -0.65
    amous
    -0.65
    BOX
    -0.65
    OTUS
    -0.64
     override
    -0.64
    POSITIVE LOGITS
     Fin
    3.92
    Fin
    2.51
     fin
    1.96
     FIN
    1.90
    fin
    1.87
    FIN
    1.31
     fins
    1.08
     finan
    1.05
     Flan
    1.04
     Finance
    1.00
    Activations Density 0.025%

    No Known Activations