INDEX

Explanations

mentions of being hired or fired

oai_token-act-pair · gpt-3.5-turbo

the word "fired" and its variations or related concepts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Bei

-0.75

alter

-0.74

adra

-0.70

 Shant

-0.68

 Gamble

-0.63

 Dunk

-0.62

doi

-0.58

åĳ

-0.58

 Wolfgang

-0.56

 Sheen

-0.56

POSITIVE LOGITS

ired

1.22

IRED

1.06

iring

1.03

ategor

0.91

ansas

0.84

ilage

0.83

ragon

0.75

reenshots

0.72

rived

0.72

irements

0.72

Activations Density 0.010%

mentions of being hired or fired

the word "fired" and its variations or related concepts

No Comments

No Known Activations