INDEX

Explanations

phrases indicating prior research and studies

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

MultipartFile

-0.59

Hentet

-0.58

 itſelf

-0.54

 getM

-0.53

ArgsConstructor

-0.53

 istore

-0.50

IntoConstraints

-0.50

TypedDataSet

-0.50

rinfo

-0.49

bells

-0.49

POSITIVE LOGITS

 previous

1.20

Previous

1.20

 Previous

1.20

previous

1.12

 previously

1.06

 PREVIOUS

1.05

PREVIOUS

0.98

 Previously

0.90

previously

0.89

Previously

0.89

Activations Density 0.157%

phrases indicating prior research and studies

No Comments

No Known Activations