INDEX

Explanations

mentions of the word "flat" and its variations in different contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

 queſta

-0.65

rungsseite

-0.61

 pouvoit

-0.61

 getNumber

-0.60

NewUrlParser

-0.58

legion

-0.58

 szerző

-0.57

 TimeUnit

-0.57

 endforeach

-0.57

SharedDtor

-0.57

POSITIVE LOGITS

flat

1.02

 FLAT

1.02

 Flat

1.02

 flat

1.01

Flat

1.01

FLAT

0.93

 flats

0.74

 Flats

0.70

flats

0.63

 flatly

0.57

Activations Density 0.005%

mentions of the word "flat" and its variations in different contexts

No Comments

No Known Activations