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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 10-RES-JB
    11. 9254
    Prev
    Next
    INDEX
    Explanations

    phrases related to commands, requests, or instructions

    oai_token-act-pair · gpt-3.5-turbo

    punctuated phrases and conjunctions that indicate significant narrative moments or transitions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    10
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    iciency
    -0.78
    mite
    -0.76
    mble
    -0.74
    rament
    -0.74
    ahime
    -0.72
    asury
    -0.71
    rastructure
    -0.68
    itizens
    -0.68
    ynam
    -0.67
    è£ħ
    -0.67
    POSITIVE LOGITS
     lest
    1.20
     prompting
    1.09
     implying
    1.09
     fearing
    1.08
     but
    1.07
     saying
    1.06
     insisting
    1.04
     joking
    1.00
     reminding
    0.99
     telling
    0.98
    Activations Density 0.393%

    No Known Activations