INDEX

Explanations

questions mentioned in news or interviews

oai_token-act-pair · gpt-3.5-turbo

the phrase "when asked," indicating inquiries or questions posed to individuals

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

mit

-0.68

Es

-0.66

ãĥĩãĤ£

-0.65

ports

-0.64

marine

-0.64

xon

-0.63

burgh

-0.62

-0.61

tten

-0.61

tiny

-0.59

POSITIVE LOGITS

 quizz

1.00

 rhet

0.90

 posed

0.83

ioned

0.79

 asked

0.78

 probing

0.73

 questions

0.70

 questioned

0.67

 confronted

0.66

erville

0.66

Activations Density 0.026%

questions mentioned in news or interviews

the phrase "when asked," indicating inquiries or questions posed to individuals

No Comments

No Known Activations