INDEX

Explanations

keywords related to news articles, specifically related to the term "feed."

oai_token-act-pair · gpt-3.5-turbo

references to social media platforms

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ultra

-0.70

 cigar

-0.69

 international

-0.66

 statute

-0.66

 constitutional

-0.65

va

-0.63

 coron

-0.63

 illegal

-0.63

 convicted

-0.62

 statues

-0.62

POSITIVE LOGITS

feed

4.50

Feed

2.63

feeding

2.10

 Feed

1.93

 feeds

1.76

 feed

1.73

fed

1.70

rss

1.26

dump

1.18

flow

1.16

Activations Density 0.010%

keywords related to news articles, specifically related to the term "feed."

references to social media platforms

No Comments

No Known Activations