INDEX

Explanations

programming-related terms, such as identifiers and method calls

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

html/java code structure

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

@}

-0.38

 Gouvernement

-0.29

天

-0.29

devtool

-0.29

 trouvez

-0.29

 Galo

-0.28

forall

-0.28

ובר

-0.28

ERVE

-0.28

kezik

-0.28

POSITIVE LOGITS

RenderAtEndOf

0.71

 kasarigan

0.71

دانشنامهٔ

0.67

+#+#

0.65

⤹

0.63

tanleria

0.62

MockMvc

0.61

parsedMessage

0.61

AISSEE

0.60

 المعيارى

0.60

Activations Density 0.012%

programming-related terms, such as identifiers and method calls

html/java code structure

No Comments

No Known Activations