INDEX

Explanations

references to places or locations

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

lounge or Loudi

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

cipar

-0.47

 fizer

-0.43

fährt

-0.40

Tet

-0.37

}))

-0.36

])))

-0.36

 eint

-0.35

Karel

-0.35

$)

-0.35

<b>

-0.35

POSITIVE LOGITS

 lounge

1.34

 Lounge

1.33

lounge

1.28

Lounge

1.23

 lounges

1.04

UNGE

0.95

 loung

0.89

 houſe

0.85

 lounging

0.82

 Houſe

0.80

Activations Density 0.002%

references to places or locations

lounge or Loudi

No Comments

No Known Activations