INDEX

Explanations

instances of the name "Robert."

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

gdx

-0.62

 initView

-0.58

 pageNo

-0.56

̈́

-0.55

detectChanges

-0.55

innis

-0.54

 useStyles

-0.54

 szüks

-0.52

swatch

-0.52

pakatan

-0.51

POSITIVE LOGITS

 Robert

0.91

 Rober

0.81

 Roberts

0.79

Robert

0.73

 Roberta

0.66

 ROBERT

0.66

 robert

0.63

 ROBERTS

0.59

robert

0.58

Roberts

0.56

Activations Density 0.005%

instances of the name "Robert."

No Comments

No Known Activations