INDEX

Explanations

mathematical notation and operational language within calculations

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

StoryboardSegue

-0.64

ंदीखरीदारी

-0.53

 transfieras

-0.52

rrggbb

-0.51

BaseAdapter

-0.47

MessageOf

-0.47

transQ

-0.47

ftagPool

-0.47

AntiForgeryToken

-0.47

KURZBESCHREIBUNG

-0.47

POSITIVE LOGITS

BorderStyle

0.44

ută

0.38

INCREF

0.37

 básica

0.36

 tiroirs

0.35

 iniziale

0.35

нятно

0.34

 THANKS

0.34

เปล

0.34

 initial

0.34

Activations Density 0.006%

mathematical notation and operational language within calculations

No Comments

No Known Activations