INDEX

Explanations

terms associated with physical harm and liability

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_0/width_16k/average_l0_105

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.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

лен

-0.56

тию

-0.55

 dessus

-0.52

 volont

-0.51

rite

-0.51

ثر

-0.51

 vald

-0.49

dov

-0.49

 sfer

-0.48

しか

-0.47

POSITIVE LOGITS

 يتيمه

0.98

 ब्रेकडाउन

0.85

 HasFactory

0.80

 متعلقه

0.79

LookAnd

0.76

 <=",

0.75

 kasarigan

0.75

 Meksiku

0.72

kloped

0.72

ecap

0.71

Activations Density 0.140%

terms associated with physical harm and liability

No Comments

No Known Activations