INDEX

Explanations

phrases indicating entering a place or starting an activity

oai_token-act-pair · gpt-3.5-turbo

repeated occurrences of the phrase "go in" with varying contexts

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

CLASSIFIED

-0.76

=]

-0.71

CVE

-0.69

STER

-0.62

Missing

-0.62

terday

-0.61

Percent

-0.61

Rest

-0.58

 ende

-0.57

%"

-0.57

POSITIVE LOGITS

ordinate

1.15

 front

1.02

humane

1.02

wards

1.01

 vitro

1.01

bred

0.98

 accordance

0.95

sole

0.95

bound

0.95

between

0.94

Activations Density 0.209%

phrases indicating entering a place or starting an activity

repeated occurrences of the phrase "go in" with varying contexts

No Comments

No Known Activations