INDEX

Explanations

words and phrases related to terrorism and violence.

oai_token-act-pair · gemini-2.0-flash

words related to terrorism and violence

np_token-act-pair-logits · gpt-4o-mini

attack/violence

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-transcoders/layer_21/width_16k/average_l0_13

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.ln2.hook_normalized

Architecture

jumprelu_transcoder

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 &___

-0.52

fø

-0.43

 uintptr

-0.43

<?>

-0.39

Portail

-0.37

DAF

-0.37

Azi

-0.36

RegressionTest

-0.36

 drin

-0.35

дат

-0.35

POSITIVE LOGITS

 attack

1.27

 bombing

1.13

 attacks

1.05

SequentialGroup

1.04

attack

1.03

 Attack

1.03

 bomb

1.01

 shooting

1.00

 ATTACK

0.99

 bombings

0.96

Activations Density 3.452%

words and phrases related to terrorism and violence.

words related to terrorism and violence

attack/violence

No Known Activations