INDEX

Explanations

words related to people's names, especially "Jer" with varying numbers following it

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Jer."

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

ULAR

-0.70

clusive

-0.68

ãģ¦

-0.68

esan

-0.67

spring

-0.64

cry

-0.63

Spread

-0.63

holder

-0.62

erella

-0.61

 screened

-0.60

POSITIVE LOGITS

seys

1.33

Jer

1.18

miah

1.09

usalem

1.06

Jer

0.97

jer

0.96

sey

0.94

jer

0.79

okers

0.77

onica

0.76

Activations Density 0.006%

words related to people's names, especially "Jer" with varying numbers following it

mentions of the name "Jer."

No Comments

No Known Activations