INDEX

Explanations

words related to politics and government

oai_token-act-pair · gpt-3.5-turbo

references to political groups and their associated figures

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

 Beaut

-0.68

 mechanic

-0.64

 conqu

-0.62

 beaut

-0.60

 cleaner

-0.58

egu

-0.57

 Mistress

-0.57

utters

-0.57

 Mama

-0.56

 mosa

-0.55

POSITIVE LOGITS

 alike

1.26

ervatives

0.93

 regarding

0.92

 condemning

0.90

 concerning

0.87

who

0.87

 concerned

0.86

 alarmed

0.84

who

0.84

 whom

0.84

Activations Density 0.385%

words related to politics and government

references to political groups and their associated figures

No Comments

No Known Activations