INDEX

Explanations

code-related terminology and testing function structures

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

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

ArgsConstructor

-0.91

 kasarigan

-0.90

exitRule

-0.85

 queſta

-0.77

aarrggbb

-0.77

تقاوى

-0.72

ंदीखरीदारी

-0.72

niſſe

-0.72

 laſſen

-0.71

VersionUID

-0.71

POSITIVE LOGITS

 window

0.30

window

0.27

XtraBars

0.27

0.26

 performance

0.26

 nodding

0.26

kok

0.26

Window

0.26

 لف

0.26

strictEqual

0.26

Activations Density 0.010%

code-related terminology and testing function structures

No Comments

No Known Activations