INDEX

Explanations

references to individuals or persons

oai_token-act-pair · gpt-3.5-turbo

references to individuals or the concept of "person."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Lans

-0.71

unctions

-0.69

NP

-0.66

 Lions

-0.65

urations

-0.64

 Lank

-0.64

Tx

-0.64

DL

-0.64

 Coliseum

-0.63

onder

-0.62

POSITIVE LOGITS

hood

1.19

nel

1.01

acles

0.83

nels

0.80

uscript

0.78

acle

0.75

ification

0.75

who

0.74

wise

0.74

ifications

0.74

Activations Density 0.035%

references to individuals or persons

references to individuals or the concept of "person."

No Comments

No Known Activations