INDEX

Explanations

references to cottage-related terms

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

types of dwellings

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

 Sudan

-0.47

↵

-0.45

 Wilson

-0.45

CMYK

-0.44

CNN

-0.43

 ASME

-0.43

 Sandra

-0.42

م

-0.42

 Rihanna

-0.42

 Lenin

-0.42

POSITIVE LOGITS

 cottage

1.22

cottage

1.16

 Cottage

1.13

 cottages

1.05

 Cottages

0.97

 cabaña

0.87

 cabin

0.81

 Cabin

0.76

parsedMessage

0.74

 cabins

0.73

Activations Density 0.006%

references to cottage-related terms

types of dwellings

No Comments

No Known Activations