INDEX

Explanations

phrases related to the workplace

oai_token-act-pair · gpt-3.5-turbo

references to workplaces or related terms

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

imaru

-0.64

eru

-0.60

 heading

-0.59

ORS

-0.59

 bursting

-0.57

aud

-0.57

 issuing

-0.57

ocytes

-0.56

 overd

-0.55

reb

-0.55

POSITIVE LOGITS

place

4.69

places

2.79

Place

2.17

placed

1.90

 place

1.76

 Place

1.70

 Places

1.22

force

1.12

 places

1.08

where

1.03

Activations Density 0.014%

phrases related to the workplace

references to workplaces or related terms

No Comments

No Known Activations