INDEX

Explanations

keywords related to specificity and individual components in various contexts

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

single or individual nouns

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

 calendriers

-0.31

 litoral

-0.31

ミアム

-0.30

 Exactos

-0.29

 Muster

-0.29

 Sachen

-0.28

 tighter

-0.28

 jupe

-0.28

 tanta

-0.28

Background

-0.28

POSITIVE LOGITS

 single

0.82

single

0.81

 individual

0.77

individual

0.73

Single

0.73

 SINGLE

0.72

 Single

0.71

 singolo

0.71

Individual

0.71

单个

0.68

Activations Density 0.073%

keywords related to specificity and individual components in various contexts

single or individual nouns

No Comments

No Known Activations