INDEX

Explanations

comparative adjectives and phrases highlighting superiority or distinction

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

"greatest common divisor/factor" or "highest common divisor/factor"

np_acts-logits-general · gemini-2.0-flash

greatest common divisor or highest levels

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

-0.63

-0.61

er

-0.59

-0.57

an

-0.57

ers

-0.57

ish

-0.56

 enough

-0.56

-0.54

POSITIVE LOGITS

 ever

1.29

 possible

1.24

possible

1.11

 EVER

1.09

 POSSIBLE

0.99

 możli

0.94

Ever

0.94

 Ever

0.93

 Possible

0.90

 imaginable

0.90

Activations Density 0.119%

comparative adjectives and phrases highlighting superiority or distinction

"greatest common divisor/factor" or "highest common divisor/factor"

greatest common divisor or highest levels

No Comments

No Known Activations