INDEX

Explanations

locations or sites mentioned in a text

oai_token-act-pair · gpt-3.5-turbo

references to locations or sites

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

âĺ

-0.76

Enlarge

-0.74

=""

-0.69

::::

-0.67

 repertoire

-0.66

ean

-0.65

/"

-0.65

wheel

-0.64

 sshd

-0.64

ophone

-0.64

POSITIVE LOGITS

 worship

0.91

 refuge

0.86

 Refuge

0.80

 holiest

0.73

 alleged

0.73

 pilgrimage

0.71

 residence

0.71

 infamous

0.70

 tallest

0.70

 showers

0.69

Activations Density 0.117%

locations or sites mentioned in a text

references to locations or sites

No Comments

No Known Activations