INDEX

Explanations

elements of discourse or conversational markers

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

 pinulongan

-0.68

rungsseite

-0.61

__":

-0.58

expandindo

-0.56

__':

-0.55

Tembelea

-0.53

 kasarigan

-0.53

 verkau

-0.52

 âgées

-0.52

dchen

-0.51

POSITIVE LOGITS

isContained

0.32

freopen

0.26

 blown

0.26

 explanation

0.25

blown

0.25

деми

0.24

sł

0.24

 onBind

0.24

 toit

0.23

æg

0.23

Activations Density 0.060%

elements of discourse or conversational markers

No Comments

No Known Activations