INDEX

Explanations

phrases or keywords related to giving directions or instructions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "Take" and variations of it in different contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 unspecified

-0.75

 remaining

-0.71

 distinguished

-0.64

 specified

-0.63

 appearing

-0.63

 both

-0.62

 otherwise

-0.62

 multiple

-0.62

 comments

-0.61

 possibly

-0.60

POSITIVE LOGITS

Take

3.23

Consider

1.83

take

1.80

Taking

1.57

Give

1.50

Put

1.43

Think

1.42

Imagine

1.41

Look

1.41

Tell

1.40

Activations Density 0.012%

phrases or keywords related to giving directions or instructions

instances of the word "Take" and variations of it in different contexts

No Comments

No Known Activations