INDEX

Explanations

HTML tags, specifically bold and strong elements

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

HTML bold tags followed by specific nouns

np_acts-logits-general · gemini-2.0-flash

break schedule

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

Glej

-0.76

<eos>

-0.73

 […]

-0.67

in

-0.60

ini

-0.60

con

-0.60

to

-0.58

 into

-0.57

POSITIVE LOGITS

 myſelf

1.22

 pleaſure

1.16

 ſtate

1.15

 itſelf

1.15

 purpoſe

1.15

 poffible

1.12

 ſmall

1.11

 Majefty

1.09

 ſeveral

1.09

</strong>

1.09

Activations Density 0.041%

HTML tags, specifically bold and strong elements

HTML bold tags followed by specific nouns

break schedule

No Comments

No Known Activations