INDEX

Explanations

symbols or indicators related to code structures or operations

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-pt-res/layer_2/width_16k/average_l0_67

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.2.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

 ſta

-0.66

 myſelf

-0.65

 Majefty

-0.64

 enfans

-0.61

 Monfieur

-0.60

 Perſ

-0.59

 Anſ

-0.59

 unſ

-0.59

 ſeveral

-0.59

 feroit

-0.58

POSITIVE LOGITS

2.67

1.30

$+$

1.29

$+

1.27

+\

1.20

＋

1.16

$+\

1.11

+=

1.06

1.05

 плюс

1.02

Activations Density 0.239%

symbols or indicators related to code structures or operations

No Comments

No Known Activations