INDEX

Explanations

occurrences of table formatting or specific structured elements resembling tabular data

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

array delimiters and math formatting

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-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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

llllll

-0.57

llllllll

-0.56

lllllll

-0.55

lllll

-0.53

lllllllll

-0.50

parsedMessage

-0.49

bootstrapcdn

-0.48

expandindo

-0.48

llll

-0.47

دانشنامهٔ

-0.47

POSITIVE LOGITS

0.56

0.52

0.50

forhold

0.49

 beginnetje

0.49

 defaultstate

0.48

 AssemblyCulture

0.47

 compactness

0.47

IContainer

0.46

icc

0.46

Activations Density 0.001%

occurrences of table formatting or specific structured elements resembling tabular data

array delimiters and math formatting

No Comments

No Known Activations