INDEX

Explanations

references to government and law enforcement agencies

oai_token-act-pair · gpt-3.5-turbo

phrases related to governmental or official institutions and their actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 contagious

-0.56

kay

-0.54

voy

-0.53

 spac

-0.52

Wave

-0.51

hol

-0.51

ï¸

-0.50

 romantic

-0.50

 spoil

-0.49

 nausea

-0.48

POSITIVE LOGITS

ngth

0.66

 headquartered

0.65

exempt

0.61

afety

0.61

incial

0.61

 overseen

0.59

 oversees

0.59

 accredited

0.57

 trustees

0.57

issued

0.56

Activations Density 1.276%

references to government and law enforcement agencies

phrases related to governmental or official institutions and their actions

No Comments

No Known Activations