INDEX

Explanations

the word "Me"

oai_token-act-pair · gpt-3.5-turbo

references to the term "Me" 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 @ 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

*/(

-0.86

paralle

-0.74

trump

-0.69

 totality

-0.66

 legality

-0.64

çĶŁ

-0.63

edged

-0.63

ordinate

-0.63

medium

-0.61

EStreamFrame

-0.61

POSITIVE LOGITS

asuring

1.00

zzo

1.00

asured

0.97

adow

0.97

ghan

0.92

eker

0.89

gan

0.89

adows

0.88

lees

0.88

asure

0.88

Activations Density 0.013%

the word "Me"

references to the term "Me" in various contexts

No Comments

No Known Activations