INDEX

Explanations

phrases or words that are being highlighted or emphasized

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "highlight" and its variations

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

 orders

-0.72

obe

-0.65

 Orders

-0.64

 disposed

-0.64

 Written

-0.62

io

-0.60

 consent

-0.60

 willingly

-0.60

 Unable

-0.59

 inhab

-0.59

POSITIVE LOGITS

 highlight

3.69

 highlights

2.54

 highlighting

1.94

 showcase

1.65

 highlighted

1.64

 Highlights

1.54

 standout

1.47

 underscore

1.41

 emphasize

1.35

 spotlight

1.24

Activations Density 0.019%

phrases or words that are being highlighted or emphasized

occurrences of the word "highlight" and its variations

No Comments

No Known Activations