INDEX

Explanations

references to small sizes or quantities

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

 filePath

-0.54

routeProvider

-0.53

__(/*!

-0.53

 CommandType

-0.52

Зноскі

-0.50

 helyzet

-0.48

 ActionListener

-0.47

 gdyby

-0.47

ActionEvent

-0.47

 linearity

-0.46

POSITIVE LOGITS

Small

1.27

 Small

1.19

 small

1.18

small

1.17

 SMALL

1.14

SMALL

1.09

 ſmall

0.93

 smal

0.89

Smaller

0.89

 Smal

0.87

Activations Density 0.117%

references to small sizes or quantities

No Comments

No Known Activations