INDEX

Explanations

sequence markers or formatting elements in a structured document

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

 resourceCulture

-0.68

 Autorizaciones

-0.65

 سكانية

-0.64

uxxxx

-0.64

styleType

-0.64

 Autorisations

-0.63

AndEndTag

-0.61

 للمعارف

-0.60

 ويكيپيديا

-0.59

 onAnimation

-0.58

POSITIVE LOGITS

WriteLiteral

0.36

hint

0.35

 feit

0.34

Update

0.33

Diss

0.33

Data

0.32

Info

0.32

🏻

0.32

Lazy

0.32

姨

0.32

Activations Density 0.003%

sequence markers or formatting elements in a structured document

No Comments

No Known Activations