INDEX

Explanations

people's names and mentions of official positions or titles

oai_token-act-pair · gpt-3.5-turbo

references to individuals' titles or positions and their associated statuses in a narrative context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 consumers

-0.70

 embryos

-0.68

 ticks

-0.66

worms

-0.66

urtle

-0.64

 deform

-0.64

 consume

-0.64

isf

-0.63

 frogs

-0.62

bender

-0.62

POSITIVE LOGITS

 lieutenant

1.48

utenant

1.28

 Lieutenant

1.27

 secretary

1.23

 Deputy

1.14

 colonel

1.14

 deputy

1.08

Lt

1.06

chief

1.05

Lt

1.05

Activations Density 0.596%

people's names and mentions of official positions or titles

references to individuals' titles or positions and their associated statuses in a narrative context

No Comments

No Known Activations