INDEX

Explanations

phrases indicating processes or methods of obtaining results

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

 transfieras

-0.46

 ennemi

-0.42

 niega

-0.40

󠁮

-0.39

fromnode

-0.38

GEBURTS

-0.37

 zieken

-0.37

новниш

-0.37

 Civilian

-0.37

curities

-0.36

POSITIVE LOGITS

Cyfarwyddwr

0.52

isMethod

0.51

 process

0.48

use

0.47

҉

0.47

 transform

0.47

 direct

0.46

 repeated

0.45

 systematically

0.45

 repeat

0.45

Activations Density 0.063%

phrases indicating processes or methods of obtaining results

No Comments

No Known Activations