INDEX

Explanations

addresses and locations indicated in text

oai_token-act-pair · gpt-3.5-turbo

address locations or geographical identifiers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

merga

-0.92

odic

-0.83

 largeDownload

-0.75

 herds

-0.72

 proteins

-0.72

 grains

-0.71

 berries

-0.65

iosis

-0.64

lied

-0.64

 grou

-0.64

POSITIVE LOGITS

 Suite

1.35

 Seattle

1.04

 Pasadena

1.02

 Dublin

1.00

 Melbourne

1.00

 Jacksonville

0.99

 Louisville

0.98

 Bangkok

0.98

 Minneapolis

0.98

 downtown

0.97

Activations Density 0.123%

addresses and locations indicated in text

address locations or geographical identifiers

No Comments

No Known Activations