INDEX

Explanations

phrases related to location or origin

oai_token-act-pair · gpt-3.5-turbo

phrases related to location or origin

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

 crow

-0.76

 eject

-0.74

 recess

-0.72

 empt

-0.71

 protr

-0.67

 shovel

-0.66

 Bohem

-0.65

 Remem

-0.64

 compartment

-0.64

 farewell

-0.64

POSITIVE LOGITS

based

3.53

Based

2.47

 based

1.95

 Based

1.76

quartered

1.66

driven

1.64

derived

1.64

focused

1.54

oriented

1.48

powered

1.45

Activations Density 0.016%

phrases related to location or origin

phrases related to location or origin

No Comments

No Known Activations