INDEX

Explanations

phrases related to interactions with other individuals

oai_token-act-pair · gpt-3.5-turbo

references to individuals or the concept of "someone" in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

DOS

-0.80

tnc

-0.72

bows

-0.68

UV

-0.67

HL

-0.67

osterone

-0.67

CDC

-0.66

ortex

-0.64

nyder

-0.62

heny

-0.60

POSITIVE LOGITS

 else

2.14

Else

1.46

else

1.45

 Else

1.36

who

1.02

WithNo

0.86

 resembling

0.82

 unfamiliar

0.81

 whom

0.81

who

0.80

Activations Density 0.068%

phrases related to interactions with other individuals

references to individuals or the concept of "someone" in various contexts

No Comments

No Known Activations