INDEX

Explanations

the repeated use of the word "to" indicating purpose or direction

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

httphttps

-0.39

))^

-0.39

TemporalType

-0.38

cô

-0.38

 nakalista

-0.37

めでとう

-0.37

庞

-0.37

olia

-0.37

stdc

-0.35

CDCl

-0.35

POSITIVE LOGITS

Wanna

0.69

 wanna

0.66

就想

0.66

 informée

0.63

 Wanna

0.63

wanna

0.63

很想

0.62

be

0.61

 WANT

0.59

 want

0.59

Activations Density 0.028%

the repeated use of the word "to" indicating purpose or direction

No Comments

No Known Activations