INDEX

Explanations

instances of dialogue or spoken words

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

Following "da", "the", or "ther"

np_acts-logits-general · gemini-2.0-flash

common articles and word beginnings

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

 betweenstory

-1.10

tvguidetime

-0.90

 мәкал

-0.87

ArgsConstructor

-0.87

ंदीखरीदारी

-0.86

 NSCoder

-0.83

Geplaatst

-0.82

findpost

-0.80

 Roskov

-0.80

+#+#

-0.80

POSITIVE LOGITS

 same

0.76

 latter

0.73

 entire

0.72

 last

0.71

 whole

0.68

 original

0.68

 main

0.66

0.62

ir

0.62

setw

0.59

Activations Density 0.092%

instances of dialogue or spoken words

Following "da", "the", or "ther"

common articles and word beginnings

No Comments

No Known Activations