INDEX

Explanations

text related to familiarity with certain topics

oai_token-act-pair · gpt-3.5-turbo

phrases indicating familiarity or unfamiliarity with a subject

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

earchers

-0.65

alez

-0.62

 teasp

-0.61

MO

-0.60

 surprised

-0.60

 tribute

-0.60

ratulations

-0.60

urches

-0.59

aughters

-0.59

emetery

-0.59

POSITIVE LOGITS

¥µ

0.86

 Norn

0.81

 firsthand

0.76

 Runes

0.72

ĳå£«

0.69

 whats

0.69

İĭ

0.67

 whom

0.67

ĺħ

0.64

 Oath

0.64

Activations Density 0.079%

text related to familiarity with certain topics

phrases indicating familiarity or unfamiliarity with a subject

No Comments

No Known Activations