INDEX

Explanations

words related to a specific city, Leicester, and possibly other cities or locations

oai_token-act-pair · gpt-3.5-turbo

references to the city of Leicester

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

Reviewer

-0.83

psc

-0.83

sch

-0.78

icts

-0.76

ramid

-0.76

swing

-0.73

irection

-0.69

anol

-0.69

DonaldTrump

-0.69

channelAvailability

-0.69

POSITIVE LOGITS

icester

1.00

 Leicester

0.86

shire

0.81

 Manor

0.79

 Tigers

0.77

wich

0.76

 Yard

0.74

 Township

0.70

ury

0.69

©¶æ¥µ

0.69

Activations Density 0.008%

words related to a specific city, Leicester, and possibly other cities or locations

references to the city of Leicester

No Comments

No Known Activations