INDEX

Explanations

proper nouns such as names of politicians, historical figures, and entities

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Configuration

jbloom/Gemma-2b-IT-Residual-Stream-SAEs/gemma_2b_it_blocks.12.hook_resid_post_16384

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

HuggingFaceFW/fineweb

Features

16,384

Data Type

float32

Hook Name

blocks.12.hook_resid_post

Hook Layer

Architecture

standard

Context Size

1,024

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 persil

-0.93

 affez

-0.90

 poiv

-0.83

 cioc

-0.80

 poliester

-0.76

 trouva

-0.76

 poire

-0.76

 ristor

-0.76

 abito

-0.73

 menthe

-0.72

POSITIVE LOGITS

 also

0.86

 furthermore

0.82

 thereupon

0.80

 impelled

0.78

 withal

0.77

 Whence

0.73

 Considerable

0.73

 Studied

0.72

 caucasian

0.72

 apportioned

0.72

Activations Density 0.243%

proper nouns such as names of politicians, historical figures, and entities

No Comments

No Known Activations