INDEX

Explanations

the word "that" indicating causality or conditionals

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

BibitemShut

-0.86



-0.78

localctx

-0.78

enterOuterAlt

-0.75

 snippetHide

-0.75

 ब्रेकडाउन

-0.71

Diweddarwch

-0.70

 &___

-0.69

Xna

-0.68

SharedCtor

-0.68

POSITIVE LOGITS

 that

0.96

 bahawa

0.73

 bahwa

0.71

 rằng

0.52

 kwamba

0.51

 because

0.49

 dass

0.48

↵

0.47

that

0.47

Because

0.46

Activations Density 0.432%

the word "that" indicating causality or conditionals

No Comments

No Known Activations