INDEX

Explanations

references to specific time frames, particularly weeks

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 year

-0.69

 Jahre

-0.67

ienze

-0.65

resultCode

-0.64

 years

-0.64

HasAnnotation

-0.63

ually

-0.63

Year

-0.60

 YEAR

-0.60

 Year

-0.59

POSITIVE LOGITS

 tonight

1.18

 today

1.07

 gestern

1.02

 yesterday

0.98

 Tonight

0.95

Tonight

0.93

today

0.93

 Yesterday

0.91

 Thursday

0.91

 вчера

0.89

Activations Density 0.134%

references to specific time frames, particularly weeks

No Comments

No Known Activations