INDEX

Explanations

words related to violence and its consequences, particularly in a humanitarian context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 17-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_17/width_65k/average_l0_125

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.17.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

 last

-0.52

昨

-0.51

 which

-0.51

last

-0.50

 recent

-0.49

iste

-0.49

Erreferentziak

-0.48

-0.47

 ويكيميديا

-0.47

PANY

-0.46

POSITIVE LOGITS

 ſeveral

1.06

 myſelf

0.98

Efq

0.98

 Jefus

0.96

 Majefty

0.95

 Diſ

0.94

 itſelf

0.93

 Anſ

0.92

 Chriftian

0.92

 Theſe

0.90

Activations Density 0.035%

words related to violence and its consequences, particularly in a humanitarian context

No Comments

No Known Activations