INDEX

Explanations

proper nouns or phrases referring to individuals

oai_token-act-pair · gpt-3.5-turbo

mentions of individuals or references to people in general

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

ules

-0.71

ornings

-0.69

ulas

-0.65

antine

-0.65

cum

-0.63

marks

-0.62

ushima

-0.62

lag

-0.61

oute

-0.61

weights

-0.60

POSITIVE LOGITS

 person

3.46

 Person

2.33

person

2.31

Person

2.08

 PERSON

2.01

 persons

1.93

guy

1.54

 woman

1.48

 Persons

1.48

 participant

1.35

Activations Density 0.023%

proper nouns or phrases referring to individuals

mentions of individuals or references to people in general

No Comments

No Known Activations