INDEX

Explanations

phrases related to actions, events, and consequences

oai_token-act-pair · gpt-3.5-turbo

terms related to calls for action or requests

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

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

Beh

-0.61

PLA

-0.59

Rating

-0.58

Ïī

-0.58

votes

-0.58

âĶĢâĶĢâĶĢâĶĢ

-0.56

Charges

-0.55

Eva

-0.55

Returns

-0.55

APE

-0.55

POSITIVE LOGITS

 abound

1.64

 vary

1.22

are

1.18

 were

1.15

 arise

1.13

 weren

1.09

 aren

1.08

 ensued

1.07

 indicate

1.07

 exist

1.05

Activations Density 0.341%

phrases related to actions, events, and consequences

terms related to calls for action or requests

No Comments

No Known Activations