INDEX

Explanations

cues indicating the start and end of structured sections in a document, such as headings or lists

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 12-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_12/width_16k/average_l0_82

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.12.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

 AUTHORS

-0.47

tak

-0.45

 trous

-0.44

yama

-0.42

app

-0.42

 MARKS

-0.41

줘

-0.41

Bal

-0.40

-0.39

setViewName

-0.39

POSITIVE LOGITS

IsContent

0.88

Personensuche

0.75

MessageOf

0.73

 дописавши

0.73

astéroïdes

0.69

 }}$}

0.69

Autoritní

0.68

RTDA

0.67

➟

0.67

=$?

0.64

Activations Density 0.233%

cues indicating the start and end of structured sections in a document, such as headings or lists

No Comments

No Known Activations