INDEX

Explanations

recurrent themes of significant words and phrases in various contexts

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

indicated that she

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

 Miß

-0.31

 cebola

-0.31

 المعيارى

-0.30

 filósofo

-0.29

 desafío

-0.28

 propOrder

-0.28

 évêque

-0.27

 pioggia

-0.27

นะนำ

-0.26

yö

-0.26

POSITIVE LOGITS

uxxxx

0.72

帖最后由

0.69

parsedMessage

0.66

HtmlAttribute

0.62

 PeEnEo

0.61

FormTagHelper

0.57

AddTagHelper

0.56

$_"

0.56

nige

0.54

ECO

0.54

Activations Density 0.040%

recurrent themes of significant words and phrases in various contexts

indicated that she

No Comments

No Known Activations