INDEX

Explanations

names and titles associated with political events or activism

oai_token-act-pair · gpt-3.5-turbo

references to individuals in a political or activist context

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

 Pike

-0.88

 bunny

-0.86

 Spokane

-0.81

 Weaver

-0.76

 nickel

-0.75

Seattle

-0.73

Columb

-0.72

TPP

-0.72

 Wilde

-0.71

wcs

-0.71

POSITIVE LOGITS

 Sheikh

1.96

 Yemeni

1.93

Abu

1.91

 Ahmed

1.90

Ibn

1.89

 Imam

1.86

 Mahmoud

1.84

 Ahmad

1.82

 Ramadan

1.80

 Mohammed

1.79

Activations Density 0.707%

names and titles associated with political events or activism

references to individuals in a political or activist context

No Comments

No Known Activations