INDEX

Explanations

instances of the word "every" followed by a number

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate proportional relationships or comparisons

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

Ü

-0.79

glers

-0.69

banks

-0.68

fried

-0.67

found

-0.64

 tongues

-0.63

ãĤ¦ãĤ¹

-0.62

 laun

-0.61

ets

-0.61

pes

-0.61

POSITIVE LOGITS

 sake

1.00

 successive

0.79

*/(

0.75

 capita

0.73

THING

0.73

 dollar

0.72

uana

0.72

kWh

0.71

 hour

0.68

resso

0.67

Activations Density 0.045%

instances of the word "every" followed by a number

phrases that indicate proportional relationships or comparisons

No Comments

No Known Activations