INDEX

Explanations

chair-related terms

oai_token-act-pair · gpt-3.5-turbo

references to positions of leadership or authority, specifically related to the term "chair" or similar variants

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

encer

-0.67

amy

-0.66

 intercepted

-0.65

 unlawfully

-0.64

 destructive

-0.62

 unintended

-0.61

imm

-0.61

 ubiqu

-0.61

 metab

-0.60

 chilled

-0.60

POSITIVE LOGITS

chair

1.77

chairs

1.24

Chair

1.24

otle

0.96

 chair

0.96

 chairs

0.95

 Chair

0.93

arten

0.90

eers

0.90

manship

0.89

Activations Density 0.005%

chair-related terms

references to positions of leadership or authority, specifically related to the term "chair" or similar variants

No Comments

No Known Activations