© 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
    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. 12895
    Prev
    Next
    INDEX
    Explanations

    the word 'inn'

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the substring "inn"

    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
     caution
    -0.64
     flies
    -0.61
    lda
    -0.61
    ============
    -0.60
    ngth
    -0.60
     Nile
    -0.60
    ffen
    -0.60
     litter
    -0.59
    gencies
    -0.58
    rely
    -0.57
    POSITIVE LOGITS
    ipeg
    1.47
    ovation
    1.23
    ikuman
    1.11
    igans
    1.02
    igan
    1.02
    umerable
    1.00
    osuke
    0.99
    ocent
    0.97
    spir
    0.96
    iband
    0.91
    Activations Density 0.010%

    No Known Activations