INDEX

Explanations

numerical data or quantifiable metrics

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

numbers and percentages

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

the

-0.60

 assist

-0.53

.~(\

-0.52

<bos>

-0.52

acters

-0.51

aid

-0.48

utter

-0.48

hdessä

-0.47

 assistance

-0.47

 Schreiber

-0.46

POSITIVE LOGITS

1.31

０

0.68

 houſe

0.67

০

0.65

۰

0.58

 XNUMX

0.54

 poffe

0.53

٠

0.53

०

0.53

 ſtate

0.53

Activations Density 0.393%

numerical data or quantifiable metrics

numbers and percentages

No Comments

No Known Activations