INDEX

Explanations

instances of the word "new."

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

new followed by various things

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

UrlResolution

-0.57

 propOrder

-0.54

 columnIndex

-0.49

enderror

-0.49

operatorname

-0.48

RemindMe

-0.47

umpang

-0.47

ISME

-0.46

 Winaray

-0.46

 동안

-0.46

POSITIVE LOGITS

new

1.40

new

1.27

NEW

1.12

 neuen

1.07

 nueva

1.07

New

1.05

 nuevos

1.05

New

1.05

NEW

1.04

 nuevas

1.04

Activations Density 0.063%

instances of the word "new."

new followed by various things

No Comments

No Known Activations