INDEX

Explanations

references to holes or openings in various contexts

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

"hole"

np_acts-logits-general · gemini-2.0-flash

hole, gap, pore

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rungsseite

-0.79

CopyWith

-0.67

AndEndTag

-0.64

Personensuche

-0.63

SequentialGroup

-0.61

RenderAtEndOf

-0.59

 للمعارف

-0.58

//});

-0.57

 initComponents

-0.57

 ivelany

-0.56

POSITIVE LOGITS

 holes

1.13

hole

1.07

 hole

1.06

 HOLE

0.94

holes

0.92

Hole

0.91

 Hole

0.91

 Holes

0.90

 openings

0.84

HOLE

0.84

Activations Density 0.345%

references to holes or openings in various contexts

"hole"

hole, gap, pore

No Comments

No Known Activations