INDEX

Explanations

quotations

oai_token-act-pair · gpt-3.5-turbo

quotation marks indicating speech or dialogue

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

 adjud

-0.72

 favour

-0.71

 appro

-0.71

tro

-0.67

 repro

-0.67

 prelim

-0.66

 lowly

-0.66

 prag

-0.66

 disappro

-0.65

 prec

-0.64

POSITIVE LOGITS

We

1.24

They

1.15

Our

1.14

It

1.13

There

1.11

1.09

Everything

1.08

Because

1.06

If

1.05

What

1.05

Activations Density 0.133%

quotations

quotation marks indicating speech or dialogue

No Comments

No Known Activations