INDEX

Explanations

words related to confirming or denying information

oai_token-act-pair · gpt-3.5-turbo

instances of the word "also" in documents

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

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

AGES

-0.71

Reviewer

-0.65

bound

-0.64

formance

-0.64

 EVEN

-0.64

ASE

-0.64

ukemia

-0.63

 srfAttach

-0.62

AGE

-0.61

Then

-0.61

POSITIVE LOGITS

 noted

1.21

 criticised

1.07

 criticized

1.07

 clarified

1.05

 cautioned

1.05

 highlighted

1.05

 discussed

1.04

 addressed

1.03

 hinted

1.02

 thanked

1.02

Activations Density 0.093%

words related to confirming or denying information

instances of the word "also" in documents

No Comments

No Known Activations