INDEX

Explanations

the phrase "what it is."

oai_token-act-pair · gpt-3.5-turbo

phrases that define or inquire about an identity or a state

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

zan

-0.69

anches

-0.65

ãĤī

-0.62

 Yard

-0.61

 Guam

-0.61

²

-0.59

visors

-0.58

oved

-0.58

tted

-0.57

mates

-0.57

POSITIVE LOGITS

 supposed

0.84

 meant

0.83

 doing

0.78

 referring

0.76

 happening

0.75

 experiencing

0.72

nt

0.70

 going

0.68

elf

0.68

 expecting

0.68

Activations Density 0.074%

the phrase "what it is."

phrases that define or inquire about an identity or a state

No Comments

No Known Activations