INDEX

Explanations

structured lists or outlines

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

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

 saites

-0.54

JspWriter

-0.53

تقاوى

-0.52

Diweddarwch

-0.49

 linkovi

-0.46

Predecesor

-0.42

LEncoder

-0.42

']>;

-0.42

resave

-0.41

AISSEE

-0.41

POSITIVE LOGITS

 كومونز

0.41

分の

0.39

keits

0.38

yon

0.35

oga

0.35

digo

0.34

 rangs

0.34

 нале

0.34

melden

0.34

ơn

0.34

Activations Density 0.053%

structured lists or outlines

No Comments

No Known Activations