INDEX

Explanations

associations related to health and well-being

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

adjective noun phrases

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

 به‌عنوان

-0.36

 dasselbe

-0.35

 enfans

-0.32

 enggak

-0.32

Ahoj

-0.31

 schoolchildren

-0.31

更好地

-0.30

yní

-0.29

 desecha

-0.29

idopsis

-0.29

POSITIVE LOGITS

 يتيمه

0.74

참고

0.69

 للمعارف

0.69

 GenerationType

0.66

etera

0.66

iblichen

0.66

 resourceCulture

0.64

ValueStyle

0.63

tagHelperRunner

0.61

脚注の使い方

0.61

Activations Density 0.158%

associations related to health and well-being

adjective noun phrases

No Comments

No Known Activations