INDEX

Explanations

instances of the word "TO" repeatedly in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 1-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_1/width_16k/average_l0_102

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.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

 présidentielle

-0.69

 bint

-0.58

 CHAPITRE

-0.57

 paff

-0.57

 Quail

-0.56

 reaſon

-0.56

 impon

-0.56

 répon

-0.56

rvats

-0.55

 becauſe

-0.55

POSITIVE LOGITS

TO

1.04

TO

0.84

ProgressHUD

0.78

FOR

0.69

ен

0.67

FOR

0.63

FROM

0.60



0.59

tou

0.58

iNdEx

0.58

Activations Density 0.035%

instances of the word "TO" repeatedly in various contexts

No Comments

No Known Activations

instances of the word "TO" repeatedly in various contexts

No Comments

No Known Activations