INDEX

Explanations

phrases related to logical reasoning or insight

oai_token-act-pair · gpt-3.5-turbo

contexts where the concept of "sense" is used to describe understanding or rationale

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

icon

-0.71

orer

-0.69

ials

-0.68

ser

-0.67

atel

-0.67

 banned

-0.65

anders

-0.63

 Recently

-0.62

url

-0.62

agged

-0.59

POSITIVE LOGITS

 sense

3.93

sense

2.46

 Sense

2.07

 senses

1.98

Sense

1.46

 perception

1.25

 semblance

1.21

 sensed

1.21

ENSE

1.21

 feeling

1.14

Activations Density 0.021%

phrases related to logical reasoning or insight

contexts where the concept of "sense" is used to describe understanding or rationale

No Comments

No Known Activations