INDEX

Explanations

engagement and interaction cues in the text

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

setVerticalGroup

-0.46

 незавершена

-0.42

AsUp

-0.42

 celo

-0.42

 navideñas

-0.41

 Absicht

-0.41

 procé

-0.40

ยิ่ง

-0.39

 Agrega

-0.39

 puncak

-0.38

POSITIVE LOGITS

########.

0.51

 reggae

0.44

पया

0.44

 oprot

0.43

squ

0.43

DIPSETTING

0.43

 sandpaper

0.42

AddTagHelper

0.42

 vampires

0.42

 Expt

0.41

Activations Density 0.028%

engagement and interaction cues in the text

No Comments

No Known Activations