INDEX

Explanations

sections that signal abrupt changes in thought or format within the text

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

Tokens after J, numbers, or November

np_acts-logits-general · gemini-2.0-flash

legal citations and case numbers

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 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

 informée

-0.81

orghini

-0.79

 témoig

-0.70

IntoConstraints

-0.66

StoryboardSegue

-0.65

Zeneca

-0.64

niſſe

-0.63

ValueStyle

-0.63

 Wikimedijinoj

-0.63

nthesis

-0.63

POSITIVE LOGITS

0.71

0.60

ChildScrollView

0.48

ly

0.44

<tr>

0.37

Pues

0.34

 éstas

0.33

 Pues

0.33

↵↵↵

0.32

</tr>

0.32

Activations Density 0.116%

sections that signal abrupt changes in thought or format within the text

Tokens after J, numbers, or November

legal citations and case numbers

No Comments

No Known Activations