INDEX

Explanations

occurrences of specific terms related to documentation, accounts, or procedures

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

before, preceding, earlier

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

sizeCache

-0.60

 AssemblyCulture

-0.58

bootstrapcdn

-0.49

NameInMap

-0.48

السكان

-0.48

ьаж

-0.48

 createSprite

-0.46

makeConstraints

-0.45

aarrggbb

-0.45

HasForeignKey

-0.44

POSITIVE LOGITS

 beforehand

0.63

 prior

0.58

 precede

0.57

 preceding

0.56

pre

0.54

 trước

0.51

 preparatory

0.50

 перед

0.50

 prep

0.50

事前

0.49

Activations Density 0.171%

occurrences of specific terms related to documentation, accounts, or procedures

before, preceding, earlier

No Comments

No Known Activations