INDEX

Explanations

mentions of specific regions or areas, particularly districts, in the context of different subjects

oai_token-act-pair · gpt-3.5-turbo

references to administrative or political districts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

BI

-0.78

Guard

-0.76

ENTION

-0.67

IDF

-0.66

ESA

-0.65

Leader

-0.65

raid

-0.65

 Liberation

-0.64

Naz

-0.64

Boy

-0.64

POSITIVE LOGITS

poons

1.22

hips

1.21

hops

1.20

ongs

1.19

uggest

1.17

paces

1.13

etter

1.09

ometimes

1.07

etting

1.06

ettings

1.05

Activations Density 0.059%

mentions of specific regions or areas, particularly districts, in the context of different subjects

references to administrative or political districts

No Comments

No Known Activations