INDEX

Explanations

negative expressions or statements indicating non-compliance or rejection

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

"If you don't" or "If not"

np_acts-logits-general · gemini-2.0-flash

if not / if don't

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 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

only

-0.66

 only

-0.63

 zwar

-0.60

 навіть

-0.59

 даже

-0.59

never

-0.57

 even

-0.57

 sekali

-0.57

なし

-0.57

even

-0.57

POSITIVE LOGITS

 already

0.91

 Already

0.77

 explicitly

0.76

 otherwise

0.74

 redan

0.73

 expressly

0.72

already

0.71

__':

0.70

 ALREADY

0.69

 immediately

0.68

Activations Density 0.219%

negative expressions or statements indicating non-compliance or rejection

"If you don't" or "If not"

if not / if don't

No Comments

No Known Activations