INDEX

Explanations

mentions of violations or violators

oai_token-act-pair · gpt-3.5-turbo

words related to violations and offenders

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

dress

-0.83

erald

-0.73

mares

-0.73

dash

-0.68

ointment

-0.67

pread

-0.67

 guiActiveUnfocused

-0.66

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

-0.63

 attrition

-0.62

iewicz

-0.62

POSITIVE LOGITS

viol

1.09

 Viol

1.03

 viol

1.00

Viol

0.88

amental

0.87

 violin

0.86

encia

0.82

ata

0.79

icious

0.78

uous

0.78

Activations Density 0.007%

mentions of violations or violators

words related to violations and offenders

No Comments

No Known Activations