INDEX

Explanations

adjectives related to negative characteristics or attitudes

oai_token-act-pair · gpt-3.5-turbo

negative descriptors related to unpleasantness or harm

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Illum

-0.74

sych

-0.73

vim

-0.71

 Emblem

-0.68

sonian

-0.67

ilot

-0.66

ĸļ

-0.65

med

-0.65

 Interpret

-0.64

 Sets

-0.64

POSITIVE LOGITS

 nasty

2.82

 degradation

2.15

 degrading

2.06

 degraded

1.95

 degrade

1.79

 vile

1.49

 vicious

1.44

 brut

1.32

 unpleasant

1.29

 horrible

1.22

Activations Density 0.033%

adjectives related to negative characteristics or attitudes

negative descriptors related to unpleasantness or harm

No Comments

No Known Activations