INDEX

Explanations

phrases describing changes in numbers or values

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

sonian

-0.67

ardi

-0.63

uminati

-0.60

sama

-0.60

ologies

-0.59

ician

-0.58

 Pastebin

-0.58

exe

-0.58

Ide

-0.56

_-

-0.56

POSITIVE LOGITS

 downhill

1.14

up

1.06

 DOWN

0.97

 backwards

0.97

 downwards

0.94

 down

0.94

 upwards

0.93

 sideways

0.90

 upward

0.88

 steadily

0.86

Activations Density 0.046%

phrases describing changes in numbers or values

No Comments

No Known Activations

phrases describing changes in numbers or values

No Comments

No Known Activations