INDEX

Explanations

verb "said" used in reported speech

oai_token-act-pair · gpt-3.5-turbo

instances of the word "said"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

otin

-0.75

à¦

-0.74

\/\/

-0.72

otion

-0.72

PHOTOS

-0.71

 ILCS

-0.70

=~=~

-0.69

Newsletter

-0.66

BUG

-0.65

EMBER

-0.65

POSITIVE LOGITS

 goodbye

0.86

 bluntly

0.80

 anecd

0.76

 sarcast

0.76

doms

0.74

 said

0.72

heit

0.72

 afterward

0.71

 afterwards

0.67

 earlier

0.66

Activations Density 0.231%

verb "said" used in reported speech

instances of the word "said"

No Comments

No Known Activations