INDEX

Explanations

names, particularly the name "Jill"

oai_token-act-pair · gpt-3.5-turbo

mentions of the name Jill

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

 eleph

-0.81

 constitu

-0.79

 tremend

-0.76

Þ

-0.76

 ingred

-0.75

ĻĤ

-0.75

rul

-0.74

culosis

-0.70

 blockers

-0.69

 tradem

-0.69

POSITIVE LOGITS

 Stein

1.19

ian

0.97

ette

0.92

ians

0.92

 Jill

0.88

annah

0.86

 Abram

0.86

 Filip

0.85

adel

0.82

athon

0.82

Activations Density 0.012%

names, particularly the name "Jill"

mentions of the name Jill

No Comments

No Known Activations