INDEX

Explanations

occurrences of section labels and mathematical notation within 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 @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

thâu

-0.82

 kasarigan

-0.79

 protoimpl

-0.77

 queſta

-0.75

OGND

-0.75

 mouseY

-0.73

دانشنامهٔ

-0.73

featureID

-0.73

ſammen

-0.72

Spolja

-0.72

POSITIVE LOGITS

section

1.42

subsection

0.80

<b>

0.80

<strong>

0.77

<h2>

0.73

<h1>

0.69

subsubsection

0.69

Section

0.68

【

0.63

mathbf

0.57

Activations Density 0.003%

occurrences of section labels and mathematical notation within the text

No Comments

No Known Activations