INDEX

Explanations

mentions of time-related phrases or recent activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 15-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_15/width_16k/average_l0_78

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.15.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

 kés

-0.48

 total

-0.46

 complete

-0.45

 asla

-0.44

 original

-0.44

шти

-0.43

reggio

-0.39

 ešte

-0.39

 rendel

-0.39

一辈子

-0.38

POSITIVE LOGITS

 lately

1.77

 Lately

1.61

recent

1.29

 recent

1.23

最近

1.20

recently

1.18

 recently

1.16

 récente

1.16

Recent

1.14

InputBorder

1.11

Activations Density 0.225%

mentions of time-related phrases or recent activities

No Comments

No Known Activations

mentions of time-related phrases or recent activities

No Comments

No Known Activations