© 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. 11-RES-JB
    6. 3892
    Prev
    Next
    INDEX
    Explanations

    phrases related to determination and assertion of personal beliefs or values

    oai_token-act-pair · gpt-3.5-turbo

    negative assertions and defiance against external pressures or expectations

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    earchers
    -0.93
    arbon
    -0.77
    idable
    -0.73
    basic
    -0.65
     aviation
    -0.64
    confirmed
    -0.60
    forestation
    -0.59
    yrinth
    -0.59
    Ear
    -0.58
    erial
    -0.58
    POSITIVE LOGITS
     condone
    1.29
     disrespect
    1.20
     apologize
    1.14
     tolerate
    1.09
     regret
    1.04
     apologise
    1.00
     hesitate
    0.96
     insult
    0.94
     resent
    0.93
     apologizing
    0.92
    Activations Density 0.289%

    No Known Activations