INDEX

Explanations

intensifiers that amplify the emotional tone of the text

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

intensifiers followed by descriptive words

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

disposing

-0.52

fromnode

-0.45

Cyfeiriadau

-0.44

 plads

-0.43

mær

-0.42

 idéia

-0.42

試試

-0.42

 olvid

-0.41

 hvid

-0.41

 dimenti

-0.41

POSITIVE LOGITS

WriteTagHelper

0.48

DeleteBehavior

0.46

OwnProperty

0.45

timewa

0.45

nodoc

0.44

parsedMessage

0.44

YourGuide

0.42

 Administrativna

0.42

TintMode

0.42

UnusedPrivate

0.42

Activations Density 0.006%

intensifiers that amplify the emotional tone of the text

intensifiers followed by descriptive words

No Comments

No Known Activations