INDEX

Explanations

phrases related to the White House and its activities

oai_token-act-pair · gpt-3.5-turbo

references to the White House

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

NetMessage

-0.78

Ö¼

-0.76

 ILCS

-0.73

eon

-0.72

#$#$

-0.72

hra

-0.71

REG

-0.71

 constitu

-0.71

utical

-0.71

ulhu

-0.70

POSITIVE LOGITS

 Correspond

1.21

 Counsel

1.14

 Office

1.04

 counsel

1.03

 Situation

0.98

 aides

0.91

 counselor

0.91

 Speech

0.87

 aide

0.84

 Press

0.84

Activations Density 0.052%

phrases related to the White House and its activities

references to the White House

No Comments

No Known Activations