INDEX

Explanations

punctuation marks, primarily periods

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

user command tasks name

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

 ſta

-0.71

 Inſ

-0.63

 ſever

-0.63

Tikang

-0.61

 BitmapFactory

-0.60

 ſche

-0.60

 diſt

-0.60

 Diſ

-0.59

 ſtate

-0.59

 faſt

-0.59

POSITIVE LOGITS

 respectivas

0.52

synthesize

0.47

 respectivos

0.46

 Formal

0.40

SerializedName

0.38

Formal

0.38

]').

0.37

 respective

0.37

$_['

0.36

 béné

0.36

Activations Density 0.007%

punctuation marks, primarily periods

user command tasks name

No Comments

No Known Activations