INDEX

Explanations

words expressing high quality or exceptional characteristics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.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

__(/*!

-0.67

 للمعارف

-0.65

GeoNames

-0.63

 '\\;'

-0.59

Portail

-0.58

Atsauces

-0.57

->___

-0.55

 насељу

-0.54

 kirj

-0.53

 Chwiliwch

-0.53

POSITIVE LOGITS

 fashion

1.14

 manner

1.02

 basis

1.00

fashion

0.91

 extent

0.90

 format

0.82

 purposes

0.80

 FASHION

0.78

basis

0.78

 biais

0.75

Activations Density 0.661%

words expressing high quality or exceptional characteristics

No Comments

No Known Activations