INDEX

Explanations

phrases indicating the location of places or items

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

located followed by spatial relation

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

parsedMessage

-0.68

 ſche

-0.65

printList

-0.64

 geſ

-0.62

 Verſ

-0.62

ntlet

-0.62

 Anſ

-0.61

 Perſ

-0.60

Accademia

-0.60

Prescription

-0.59

POSITIVE LOGITS

 located

1.20

located

1.03

 Located

1.02

 locate

1.01

Located

1.00

 Locate

0.95

 locating

0.88

Locate

0.86

 situated

0.86

 situés

0.82

Activations Density 0.009%

phrases indicating the location of places or items

located followed by spatial relation

No Comments

No Known Activations