INDEX

Explanations

references to government officials, legal terms, and organizations

oai_token-act-pair · gpt-3.5-turbo

elements related to governmental or institutional context and activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Whitman

-0.81

Defense

-0.79

avored

-0.74

 favored

-0.74

caliber

-0.73

GOP

-0.72

 Hawai

-0.69

rollment

-0.69

Minnesota

-0.68

 honoring

-0.68

POSITIVE LOGITS

 recognise

1.29

 realise

1.25

 Â£

1.23

 centre

1.22

 apologise

1.19

 apologised

1.15

 Labour

1.14

 analyse

1.14

 criticised

1.13

 aluminium

1.13

Activations Density 4.304%

references to government officials, legal terms, and organizations

elements related to governmental or institutional context and activities

No Comments

No Known Activations