INDEX

Explanations

mentions of school grades, especially emphasizing scores like grade 9 and grade 10

oai_token-act-pair · gpt-3.5-turbo

references to educational grades or performance levels

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

nect

-0.76

Roaming

-0.75

aun

-0.73

immer

-0.69

 Cele

-0.65

 Events

-0.64

eston

-0.63

 Reborn

-0.63

hots

-0.62

à¼

-0.62

POSITIVE LOGITS

 grade

4.18

 Grade

3.22

 grades

2.94

grade

2.70

Grade

2.63

 grading

2.12

 graded

2.06

grades

1.45

GPA

1.44

graded

1.28

Activations Density 0.016%

mentions of school grades, especially emphasizing scores like grade 9 and grade 10

references to educational grades or performance levels

No Comments

No Known Activations