INDEX

Explanations

programming-related terminology, particularly involving code constructs and data management

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

variable assignments, numbers, and symbols

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

 nahilalakip

-0.92

 EconPapers

-0.72

AndEndTag

-0.71

تقاوى

-0.70

 wireType

-0.69

transQ

-0.69

MLLoader

-0.63

 protoimpl

-0.63

UpInside

-0.63

 BoxDecoration

-0.61

POSITIVE LOGITS

 rather

0.38

richTextPanel

0.32

 ultimate

0.31

 always

0.30

 high

0.30

 instead

0.29

 Rather

0.29

 zwaar

0.29

bu

0.29

mdash

0.29

Activations Density 0.166%

programming-related terminology, particularly involving code constructs and data management

variable assignments, numbers, and symbols

No Comments

No Known Activations