INDEX

Explanations

phrases or words related to negative sentiments and biases

oai_token-act-pair · gpt-3.5-turbo

phrases or terms related to anti-Semitic sentiments and actions

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

 scratch

-0.89

 laure

-0.80

 fold

-0.76

 precincts

-0.74

 curled

-0.74

 gorge

-0.74

 thumbnail

-0.72

 dots

-0.72

 Brus

-0.70

 dashed

-0.70

POSITIVE LOGITS

Semitic

1.81

Semitism

1.69

immigrant

1.62

establishment

1.57

government

1.57

democratic

1.55

choice

1.53

capitalist

1.52

abortion

1.51

gay

1.51

Activations Density 0.033%

phrases or words related to negative sentiments and biases

phrases or terms related to anti-Semitic sentiments and actions

No Comments

No Known Activations