INDEX

Explanations

names, specifically the name "Julia" with a high activation value of 9 or 10

oai_token-act-pair · gpt-3.5-turbo

the mention of specific names, particularly variations of the name "Julia" and "Julie."

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

holders

-0.84

holder

-0.81

ngth

-0.80

fare

-0.71

roleum

-0.70

ribute

-0.70

enei

-0.70

rollers

-0.70

dash

-0.70

creen

-0.69

POSITIVE LOGITS

udeau

0.79

Ç

0.78

IRO

0.76

alan

0.75

 Hacker

0.75

iets

0.71

inx

0.71

 Augusta

0.71

vm

0.70

 Julia

0.70

Activations Density 0.014%

names, specifically the name "Julia" with a high activation value of 9 or 10

the mention of specific names, particularly variations of the name "Julia" and "Julie."

No Comments

No Known Activations