INDEX

Explanations

phrases related to transitions or sequence ordering

oai_token-act-pair · gpt-3.5-turbo

instances of the word "Next."

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

 Feldman

-0.78

enance

-0.68

utic

-0.65

 Sinai

-0.64

lees

-0.63

ux

-0.63

 Ãĸ

-0.63

 Coulter

-0.61

 chloride

-0.60

lee

-0.60

POSITIVE LOGITS

STEP

0.88

 generation

0.88

 Steps

0.86

door

0.85

 generations

0.80

 Generations

0.80

 Generation

0.80

 installment

0.79

Gen

0.76

 steps

0.75

Activations Density 0.034%

phrases related to transitions or sequence ordering

instances of the word "Next."

No Comments

No Known Activations