INDEX

Explanations

references to water

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

water and 'based'

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

 impressão

-0.47

Síguenos

-0.47

richTextPanel

-0.47

 crece

-0.45

참고

-0.44

 colgar

-0.44

 quedado

-0.44

 réussir

-0.44

 allá

-0.43

antula

-0.43

POSITIVE LOGITS

 water

1.18

 Water

0.98

water

0.96

Water

0.91

 WATER

0.89

 agua

0.83

の水

0.75

 Вода

0.75

 água

0.75

 waters

0.74

Activations Density 0.051%

references to water

water and 'based'

No Comments

No Known Activations