INDEX

Explanations

phrases that indicate a function or role

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

تقاوى

-0.76

expandindo

-0.73

 Taktlose

-0.62

 enqu

-0.60

 enem

-0.60

rungsseite

-0.59

migrationBuilder

-0.57

Vidite

-0.54

JvmStatic

-0.54

 beginnetje

-0.54

POSITIVE LOGITS

 toimi

0.84

起到

0.65

RunAsync

0.65

inghouse

0.63

 PROVIDED

0.61

ývá

0.58

 provided

0.58

ctiv

0.58

 przewod

0.57

 provide

0.57

Activations Density 0.198%

phrases that indicate a function or role

No Comments

No Known Activations