INDEX

Explanations

phrases or terms associated with images or picture files

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

 Halsey

-0.57

alsey

-0.55

 Sutter

-0.52

 FontWeight

-0.49

hemp

-0.48

 Dull

-0.47

 validators

-0.47

strata

-0.47

 szerző

-0.46

 mense

-0.46

POSITIVE LOGITS

Pic

0.97

Pic

0.96

PIC

0.92

 pics

0.88

PIC

0.87

pic

0.85

pic

0.84

 Pics

0.82

Pics

0.80

Picture

0.71

Activations Density 0.003%

phrases or terms associated with images or picture files

No Comments

No Known Activations