INDEX

Explanations

references to specific locations or establishments

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

place and places

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

 active

-0.46

 znan

-0.43

 avances

-0.43

 Asunción

-0.42

 Kompet

-0.42

 gydy

-0.41

 явля

-0.41

 errors

-0.40

 tüm

-0.40

 nakalista

-0.40

POSITIVE LOGITS

 stuff

0.58

 guys

0.55

 PLACE

0.54

guy

0.54

 Place

0.53

place

0.52

 Guys

0.51

 thingy

0.51

Stuff

0.50

places

0.50

Activations Density 0.007%

references to specific locations or establishments

place and places

No Comments

No Known Activations