INDEX

Explanations

terms related to covering something

oai_token-act-pair · gpt-3.5-turbo

references to something being "covered," suggesting a focus on the notion of coverage in various 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

oner

-0.70

ager

-0.70

beta

-0.69

ject

-0.68

 Alpha

-0.67

Fr

-0.65

reality

-0.65

ragon

-0.65

acting

-0.64

ithmetic

-0.63

POSITIVE LOGITS

 covered

3.71

covered

2.59

 COVER

1.89

 covers

1.85

 covering

1.82

 cover

1.82

overed

1.55

 coated

1.52

 coverage

1.47

 uncovered

1.46

Activations Density 0.019%

terms related to covering something

references to something being "covered," suggesting a focus on the notion of coverage in various contexts

No Comments

No Known Activations