INDEX

Explanations

words related to episodes of a series or program

oai_token-act-pair · gpt-3.5-turbo

references to episodes in a series

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

LESS

-0.77

IFIC

-0.76

FUL

-0.75

 Pathfinder

-0.74

WAYS

-0.67

^^^^

-0.66

DEN

-0.65

 Petty

-0.64

IFIED

-0.64

ãĥīãĥ©ãĤ´ãĥ³

-0.63

POSITIVE LOGITS

iphany

1.10

isodes

1.10

hemer

1.05

istle

1.01

iscopal

1.00

Ep

0.99

Ep

0.96

anz

0.96

igen

0.91

ais

0.88

Activations Density 0.008%

words related to episodes of a series or program

references to episodes in a series

No Comments

No Known Activations