INDEX

Explanations

questions, particularly those starting with "What"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.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

AddTagHelper

-0.57

 '\\;'

-0.57

posedge

-0.56

OGND

-0.55

 فريبيس

-0.53

Personendaten

-0.53

NameInMap

-0.52

tagHelper

-0.50

 ujednoznacz

-0.50

AddWithValue

-0.50

POSITIVE LOGITS

do

0.67

 does

0.65

 happens

0.57

are

0.57

 Does

0.52

 Happens

0.52

 about

0.51

Does

0.48

 follows

0.47

 happened

0.46

Activations Density 0.121%

questions, particularly those starting with "What"

No Comments

No Known Activations