INDEX

Explanations

references to numerical values

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

geries

-0.62

ãĥ¼ãĥĨãĤ£

-0.60

inventoryQuantity

-0.58

 vanity

-0.57

 booted

-0.57

nob

-0.56

itures

-0.56

GUN

-0.55

 Kills

-0.55

ãĥŀ

-0.55

POSITIVE LOGITS

].

1.18

–

1.14

â

1.12

],[

1.12

].

1.10

]).

1.01

âĢĲ

1.00

âĢ¢âĢ¢

0.96

]),

0.96

reviewed

0.93

Activations Density 0.059%

references to numerical values

No Comments

No Known Activations