INDEX

Explanations

instances of the word "lie" in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_0/width_16k/average_l0_105

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Meksiku

-0.72

AndEndTag

-0.70

 snowball

-0.70

 Guin

-0.68

LEncoder

-0.63

Hez

-0.62

 Fasc

-0.62

 AspNetCore

-0.62

fällig

-0.62

jména

-0.62

POSITIVE LOGITS

 lies

1.63

lie

1.59

sat

1.50

 sits

1.48

sit

1.48

 lying

1.28

 sitting

1.24

 Lying

1.13

Sitting

1.07

 Sitting

1.06

Activations Density 0.088%

instances of the word "lie" in various contexts

No Comments

No Known Activations