INDEX

Explanations

terms associated with verifying information and critical analysis

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 enfance

-0.43

 FetchType

-0.42

 plazos

-0.38

Хро

-0.38

KURZBESCHREIBUNG

-0.36

 parís

-0.36

 comunión

-0.35

eseorang

-0.35

ikistan

-0.35

 tetrachloride

-0.34

POSITIVE LOGITS

 students

0.75

Students

0.69

 Students

0.68

 student

0.60

students

0.59

 STUDENTS

0.58

 Student

0.56

Student

0.55

 STUDENT

0.55

 estudiantes

0.55

Activations Density 0.194%

terms associated with verifying information and critical analysis

No Comments

No Known Activations