INDEX

Explanations

references to holes or openings in various contexts

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

"hole" or "holes"

np_acts-logits-general · gemini-2.0-flash

holes and openings

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

-0.91

Personensuche

-0.72

uxxxx

-0.58

 joga

-0.58

Bier

-0.54

řevě

-0.54

anyahu

-0.53

uyện

-0.53

Carex

-0.52

getSystemService

-0.52

POSITIVE LOGITS

 hole

1.46

 holes

1.42

 Hole

1.28

hole

1.27

 HOLE

1.22

Hole

1.19

 Holes

1.18

 openings

1.13

holes

1.12

Holes

1.07

Activations Density 0.402%

references to holes or openings in various contexts

"hole" or "holes"

holes and openings

No Comments

No Known Activations