INDEX

Explanations

unusual or unique words and names, especially those ending in 'on' or 'og'

oai_token-act-pair · gpt-3.5-turbo

specific names or terms related to cultural or local significance

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

 nationally

-0.71

 infringing

-0.69

 necessities

-0.68

 scratch

-0.66

 Millenn

-0.65

 worn

-0.64

 disgr

-0.63

 sorely

-0.62

 commute

-0.62

 grooming

-0.62

POSITIVE LOGITS

icz

1.13

cia

1.07

ova

1.03

ui

1.01

gian

1.01

je

1.01

inar

1.00

cs

1.00

olor

0.97

yk

0.97

Activations Density 0.346%

unusual or unique words and names, especially those ending in 'on' or 'og'

specific names or terms related to cultural or local significance

No Comments

No Known Activations