INDEX

Explanations

expressions of frustration or disappointment

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

"too"

np_acts-logits-general · gemini-2.0-flash

excess with qualifiers

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

 ModelExpression

-0.81

#+#

-0.66

FTFY

-0.62

principalTable

-0.62

abestanden

-0.62

AndEndTag

-0.57

PropertyChanging

-0.56

esModule

-0.55

enderror

-0.55

ագրություններ

-0.55

POSITIVE LOGITS

too

1.06

too

1.04

Too

1.01

Too

0.93

TOO

0.91

 demasi

0.78

TOO

0.77

 demasiado

0.76

 terlalu

0.75

 слишком

0.73

Activations Density 0.159%

expressions of frustration or disappointment

"too"

excess with qualifiers

No Comments

No Known Activations