INDEX

Explanations

patterns or structures related to data encoding or indexing

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

'1' within curly braces and colons

np_acts-logits-general · gemini-2.0-flash

structured data blocks

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

"}

-0.32

).}

-0.31

 rather

-0.31

ValueStyle

-0.31

Rohy

-0.30

richTextPanel

-0.30

 industry

-0.29

"]]

-0.29

TagMode

-0.29

gulp

-0.28

POSITIVE LOGITS

UIControlState

0.69

 surla

0.60

0.58

 Vikipedi

0.57

 dezelve

0.56

ItemBackground

0.55

 tenisky

0.54

intios

0.52

Diweddarwch

0.52

 GenerationType

0.51

Activations Density 1.126%

patterns or structures related to data encoding or indexing

'1' within curly braces and colons

structured data blocks

No Comments

No Known Activations