INDEX

Explanations

references to figures or tables in the text

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

preceding figure/Fig

np_acts-logits-general · gemini-2.0-flash

"in fig." or "in FIGURE"

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 بيها

-0.49

 își

-0.49

 palha

-0.44

 sooner

-0.43

 fédé

-0.43

 entière

-0.42

្

-0.42

 sella

-0.42

Lo

-0.41

あって

-0.40

POSITIVE LOGITS

setVerticalGroup

0.87

 >=",

0.79

 tartalomajánló

0.76

دانشنامهٔ

0.73

 kasarigan

0.70

WithMany

0.69

esgue

0.69

']);

0.69

 bezeichneter

0.69

DebuggerNonUser

0.68

Activations Density 0.506%

references to figures or tables in the text

preceding figure/Fig

"in fig." or "in FIGURE"

No Comments

No Known Activations