INDEX

Explanations

intensifiers or words expressing high degree or emphasis

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 12-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_12/width_16k/average_l0_82

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.12.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

ded

-0.69

IsContent

-0.68

 Jackman

-0.67

 Arxivat

-0.66

 проще

-0.66

klart

-0.66

 Gillette

-0.65

Griffin

-0.64

 Akhtar

-0.64

id

-0.63

POSITIVE LOGITS

 very

1.20

 Very

1.18

Very

1.17

VERY

1.16

very

1.16

 VERY

1.15

sehr

0.93

Molto

0.90

muy

0.87

Muy

0.86

Activations Density 0.057%

intensifiers or words expressing high degree or emphasis

No Comments

No Known Activations