INDEX

Explanations

references to nobility or titled individuals

oai_token-act-pair · gpt-3.5-turbo

the title "Lord" in various contexts

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

ickr

-0.80

urses

-0.79

insula

-0.74

¥ŀ

-0.73

OPLE

-0.72

urrency

-0.71

actionDate

-0.71

iences

-0.67

obos

-0.67

ank

-0.66

POSITIVE LOGITS

 Vader

0.92

 Voldemort

0.85

ess

0.84

lord

0.83

 Jesus

0.82

Lord

0.80

stown

0.80

 Krishna

0.77

 Lords

0.77

bridge

0.76

Activations Density 0.014%

references to nobility or titled individuals

the title "Lord" in various contexts

No Comments

No Known Activations