INDEX

Explanations

phrases related to political figures and events

oai_token-act-pair · gpt-3.5-turbo

phrases or statements that imply accountability or consequences

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

 glim

-0.97

 beard

-0.88

 Corsair

-0.88

beard

-0.82

 Shogun

-0.82

inburgh

-0.79

kan

-0.78

 Anaheim

-0.78

Aki

-0.77

 samurai

-0.77

POSITIVE LOGITS

 Clinton

2.37

Clinton

2.37

 Clintons

2.28

Hillary

1.93

 Hillary

1.92

clinton

1.91

 Rodham

1.81

INTON

1.77

 Abedin

1.68

hillary

1.62

Activations Density 0.449%

phrases related to political figures and events

phrases or statements that imply accountability or consequences

No Comments

No Known Activations