INDEX

Explanations

quotations or reported speech

oai_token-act-pair · gpt-3.5-turbo

quoted statements or dialogue in a narrative context

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

orno

-0.74

 desper

-0.71

umerable

-0.70

cient

-0.69

 underestimated

-0.68

Ń·

-0.67

âĦ¢:

-0.67

otin

-0.67

ħĭ

-0.65

NetMessage

-0.65

POSITIVE LOGITS

 spokeswoman

1.44

 spokesperson

1.43

 spokesman

1.37

 Acting

1.01

 spokes

0.97

 statement

0.93

 Superintendent

0.93

CEO

0.91

 Capt

0.89

 Assistant

0.87

Activations Density 0.209%

quotations or reported speech

quoted statements or dialogue in a narrative context

No Comments

No Known Activations