INDEX

Explanations

programming-related terms and data types

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

帖最后由

-0.89



-0.75

 rhestr

-0.75

IsMutable

-0.73

Datuak

-0.70

 ویکی‌پدی

-0.69

تقاوى

-0.68

AnchorStyles

-0.68

 snippetHide

-0.67

aarrggbb

-0.67

POSITIVE LOGITS

0.32

に対し

0.30

mål

0.29

 judiciaire

0.29

 whose

0.29

or

0.29

рино

0.29

paymentMethod

0.28

 guarn

0.28

 fons

0.28

Activations Density 0.019%

programming-related terms and data types

No Comments

No Known Activations