INDEX

Explanations

references to traditional methodologies or practices

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

traditional modifying various concepts

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

 Inscrivez

-0.48

 Verordnung

-0.47

 peines

-0.46

väg

-0.45

 Bewußt

-0.44

 händer

-0.44

 Durchmesser

-0.43

 varandra

-0.43

 pandemia

-0.43

lighet

-0.43

POSITIVE LOGITS

 traditional

1.11

traditional

1.09

 Traditional

1.06

Traditional

1.05

 conventional

0.96

conventional

0.93

Conventional

0.91

 Conventional

0.87

 classic

0.80

 classical

0.80

Activations Density 0.014%

references to traditional methodologies or practices

traditional modifying various concepts

No Comments

No Known Activations