INDEX

Explanations

phrases that signal conclusions, notes, or summaries in the text

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

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

ambién

-0.47

caneca

-0.42

 liền

-0.40

 Vasquez

-0.38

 concurrently

-0.37

lappe

-0.36

insatz

-0.36

 souci

-0.35

uttavia

-0.35

Hi

-0.35

POSITIVE LOGITS



0.56

 Lastly

0.55

rungsseite

0.55

featureID

0.54

fjspx

0.53

Datuak

0.53

 protoimpl

0.52

bootstrapcdn

0.52

Portail

0.52

最後は

0.52

Activations Density 0.097%

phrases that signal conclusions, notes, or summaries in the text

No Comments

No Known Activations