INDEX

Explanations

phrases indicating a lack of interest or effort

oai_token-act-pair · gpt-3.5-turbo

the term "bother" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Blueprint

-0.79

rams

-0.74

arta

-0.70

uties

-0.67

RNA

-0.67

reed

-0.66

rights

-0.62

eous

-0.62

edes

-0.62

hex

-0.60

POSITIVE LOGITS

 bother

1.49

 bothering

1.29

 bothered

0.99

 bothers

0.92

some

0.79

 worry

0.76

alore

0.73

 Grassley

0.72

swer

0.71

ername

0.69

Activations Density 0.005%

phrases indicating a lack of interest or effort

the term "bother" in various contexts

No Comments

No Known Activations