INDEX

Explanations

proper names, particularly the name "Caroline"

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals, particularly those named Caroline and Sil

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

akening

-0.91

awaru

-0.84

raising

-0.82

gettable

-0.80

riors

-0.79

unin

-0.74

arily

-0.74

ortality

-0.69

rior

-0.67

blance

-0.67

POSITIVE LOGITS

lihood

0.91

icone

0.90

enced

0.80

Jav

0.76

bury

0.73

jad

0.72

 Breed

0.71

————

0.68

coni

0.68

 Pour

0.66

Activations Density 0.043%

proper names, particularly the name "Caroline"

references to specific individuals, particularly those named Caroline and Sil

No Comments

No Known Activations