INDEX

Explanations

mentions of biased reporting

oai_token-act-pair · gpt-3.5-turbo

terms related to bias and impartiality in media reporting

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Poss

-0.73

 Survive

-0.71

 grace

-0.69

Â§

-0.66

sis

-0.65

Ã¨re

-0.64

iens

-0.63

 escape

-0.63

 Forever

-0.63

ft

-0.62

POSITIVE LOGITS

 biased

3.70

biased

2.50

 unbiased

2.32

 skewed

1.74

 impartial

1.71

 biases

1.31

 tilted

1.22

 prejud

1.21

 discriminatory

1.21

 rigged

1.16

Activations Density 0.030%

mentions of biased reporting

terms related to bias and impartiality in media reporting

No Comments

No Known Activations