INDEX

Explanations

names or references related to a specific sports team or organization, Juventus

oai_token-act-pair · gpt-3.5-turbo

mentions of the football club Juventus

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

cast

-0.72

 polar

-0.72

 stranded

-0.70

arium

-0.69

 stacking

-0.68

arrow

-0.66

banks

-0.66

aced

-0.66

 mapping

-0.65

arge

-0.65

POSITIVE LOGITS

Ju

3.90

Ju

3.10

ju

2.03

ju

1.64

 Juventus

1.53

 Juice

1.46

Jal

1.40

 Juven

1.29

Lup

1.19

 Jagu

1.14

Activations Density 0.029%

names or references related to a specific sports team or organization, Juventus

mentions of the football club Juventus

No Comments

No Known Activations