INDEX

Explanations

words related to negative or critical attitudes towards someone or something

oai_token-act-pair · gpt-3.5-turbo

terms expressing negative attitudes towards people or concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 hemor

-0.75

icle

-0.68

 Lans

-0.67

 destruct

-0.66

 helicop

-0.63

icles

-0.61

raine

-0.60

ramid

-0.60

 Staten

-0.59

 reorgan

-0.58

POSITIVE LOGITS

acy

0.86

FUL

0.85

fully

0.83

chery

0.82

fulness

0.81

uous

0.79

ately

0.78

ability

0.78

lessly

0.76

rence

0.76

Activations Density 0.040%

words related to negative or critical attitudes towards someone or something

terms expressing negative attitudes towards people or concepts

No Comments

No Known Activations