INDEX

Explanations

words and phrases indicating criticism and evaluation of events, performances, or qualities

oai_token-act-pair · gpt-4o-mini Triggered by @bot

Tokens after the word "user"

np_acts-logits-general · gemini-2.0-flash

contends that, alienation of

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 colgantes

-0.41

 Kedua

-0.37

 Gemeins

-0.36

TagMode

-0.36

 sää

-0.35

ingroup

-0.34

vzd

-0.34

/***/

-0.34

省市镇

-0.33

}])

-0.33

POSITIVE LOGITS

#+#

0.63

 estekak

0.55

UnusedPrivate

0.53

 оригіналу

0.50

ьаж

0.49

 препратки

0.49

Personensuche

0.48

 Administrativna

0.46

GTCX

0.45

cency

0.45

Activations Density 0.839%

words and phrases indicating criticism and evaluation of events, performances, or qualities

Tokens after the word "user"

contends that, alienation of

No Comments

No Known Activations