INDEX

Explanations

statements or phrases where someone is being questioned or doubted

oai_token-act-pair · gpt-3.5-turbo

instances of the word "questioned."

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

tern

-0.61

 arrangement

-0.60

 bandwidth

-0.59

AMI

-0.59

-0.58

 layout

-0.58

 display

-0.58

 placement

-0.57

 Animation

-0.57

 sync

-0.56

POSITIVE LOGITS

 questioned

3.44

 interrogated

2.02

 doubted

1.97

 questioning

1.78

 challenged

1.67

 interviewed

1.64

 criticized

1.56

 inquired

1.56

 criticised

1.55

 asked

1.54

Activations Density 0.013%

statements or phrases where someone is being questioned or doubted

instances of the word "questioned."

No Comments

No Known Activations