INDEX

Explanations

references to safety, improvement, and applications in various contexts

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

used as examples or equivalents

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 31-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_31/width_131k/average_l0_109

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.31.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

nextInt

-0.32

 respectivas

-0.32

 neither

-0.31

 betreft

-0.31

ellemző

-0.30

 voors

-0.30

two

-0.30

 suom

-0.30

 thereupon

-0.29

IsUnicode

-0.29

POSITIVE LOGITS

 ingredient

0.60

 underlie

0.59

+#+#

0.59

 sebagai

0.58

 الحره

0.57

 Ingredient

0.57

sebagai

0.56

 assisting

0.56

 jako

0.56

Used

0.56

Activations Density 0.074%

references to safety, improvement, and applications in various contexts

used as examples or equivalents

No Comments

No Known Activations