INDEX

Explanations

personal names

oai_token-act-pair · gpt-3.5-turbo

proper nouns, specifically names of individuals and places

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

âĶĢâĶĢ

-0.71

Nex

-0.68

 daytime

-0.68

 schedule

-0.63

 profit

-0.63

alt

-0.62

 heel

-0.61

 equivalent

-0.60

 increment

-0.59

 volume

-0.59

POSITIVE LOGITS

baugh

1.33

ingham

1.21

love

1.20

shall

1.16

heimer

1.12

idge

1.11

alter

1.09

ley

1.07

enberg

1.07

inski

1.06

Activations Density 0.263%

personal names

proper nouns, specifically names of individuals and places

No Comments

No Known Activations