INDEX

Explanations

instances where the text prompts the user to take action, specifically to provide an email address and subscribe

oai_token-act-pair · gpt-3.5-turbo

second-person pronouns addressing the reader directly

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

abama

-0.74

 Aberdeen

-0.70

amaz

-0.66

prim

-0.65

ÂłÂłÂłÂł

-0.64

achable

-0.63

 Aber

-0.62

iard

-0.62

 Prediction

-0.61

itism

-0.61

POSITIVE LOGITS

're

1.29

've

1.10

'll

1.09

may

0.92

'd

0.91

can

0.83

 might

0.80

 cannot

0.78

RS

0.75

 must

0.75

Activations Density 0.046%

instances where the text prompts the user to take action, specifically to provide an email address and subscribe

second-person pronouns addressing the reader directly

No Comments

No Known Activations