INDEX

Explanations

structured data or attributes in a coded format

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

The number five

np_acts-logits-general · gemini-2.0-flash

word five

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Third

-0.67

Third

-0.61

 Wednesday

-0.56

opress

-0.52

 tercera

-0.49

 third

-0.48

第三

-0.47

 Tuesday

-0.47

 Dienstag

-0.46

 WEDNESDAY

-0.46

POSITIVE LOGITS

 fifth

1.24

1.13

 Fifth

1.08

⑤

1.02

Fifth

0.99

 FIFTH

0.99

 five

0.96

 fün

0.96

fifth

0.95

第五

0.94

Activations Density 0.398%

structured data or attributes in a coded format

The number five

word five

No Comments

No Known Activations