INDEX

Explanations

instances of the word "bias" or words related to biased behavior

oai_token-act-pair · gpt-3.5-turbo

references to bias and prejudice

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

URE

-0.72

ilk

-0.72

esc

-0.70

zech

-0.70

WithNo

-0.70

ures

-0.68

Phones

-0.68

bridge

-0.67

 Danger

-0.67

Money

-0.66

POSITIVE LOGITS

 favoring

1.23

 towards

1.10

 toward

1.07

 against

1.04

icial

1.01

 prejud

0.99

 bias

0.98

 biased

0.96

 biases

0.92

 favor

0.88

Activations Density 0.033%

instances of the word "bias" or words related to biased behavior

references to bias and prejudice

No Comments

No Known Activations