INDEX

Explanations

geographical features and landmarks

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

mountains and hills

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

ambassade

-0.36

 danni

-0.35

HandlerContext

-0.34

 suun

-0.34

ıll

-0.33

kém

-0.33

ospels

-0.32

océan

-0.32

irot

-0.32

útbol

-0.31

POSITIVE LOGITS

 surla

0.80

 mountain

0.69

 Mountain

0.66

 MOUNTAIN

0.62

Mountain

0.61

expandindo

0.59

RegressionTest

0.58

 PeEnEo

0.58

 hill

0.57

 فريبيس

0.57

Activations Density 0.028%

geographical features and landmarks

mountains and hills

No Comments

No Known Activations