INDEX

Explanations

information related to organizations, numbers, and statistics

oai_token-act-pair · gpt-3.5-turbo

references to organizations and their membership statistics

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

 rumours

-0.98

 Rohingya

-0.96

 Rohing

-0.96

 Houth

-0.87

 Assange

-0.87

Nusra

-0.84

 Kejriwal

-0.84

 Corbyn

-0.84

 Boko

-0.82

 Farage

-0.82

POSITIVE LOGITS

 federally

0.94

 federal

0.85

 Federal

0.83

[[

0.82

 American

0.82

 organization

0.81

 interstate

0.80

 automobile

0.79

 governmental

0.79

 workplace

0.77

Activations Density 0.697%

information related to organizations, numbers, and statistics

references to organizations and their membership statistics

No Comments

No Known Activations