INDEX

Explanations

the word "the" when followed by short phrases

oai_token-act-pair · gpt-3.5-turbo

instances of the word "the."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 instead

-0.76

Layer

-0.74

zai

-0.74

.<

-0.71

worn

-0.71

 whilst

-0.71

 whereas

-0.70

.</

-0.68

rade

-0.68

 because

-0.68

POSITIVE LOGITS

 aforementioned

1.17

 latter

1.11

 latest

0.94

ses

0.89

 slightest

0.88

 same

0.88

 largest

0.86

 greatest

0.85

oret

0.85

 entire

0.84

Activations Density 0.684%

the word "the" when followed by short phrases

instances of the word "the."

No Comments

No Known Activations