INDEX

Explanations

references to experts and researchers involved in medical studies or investigations

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

who/which followed by action or role

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

 humanidad

-0.35

 légitime

-0.33

 borracha

-0.33

 varandra

-0.32

 skolan

-0.32

 verdaderos

-0.31

 hospod

-0.30

 vägen

-0.29

 verdadera

-0.29

 Universitaria

-0.29

POSITIVE LOGITS

OGND

0.87

 Савезне

0.80

WebServlet

0.77

ſcher

0.74

quelize

0.73

Personendaten

0.69

IVEREF

0.69

DECREF

0.68

 EconPapers

0.68

(!__

0.67

Activations Density 0.021%

references to experts and researchers involved in medical studies or investigations

who/which followed by action or role

No Comments

No Known Activations