INDEX

Explanations

names of specific locations or cities

oai_token-act-pair · gpt-3.5-turbo

mentions of locations, particularly those in Beverly Hills and Pasadena

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

emic

-0.98

*/(

-0.91

ebook

-0.90

umbn

-0.80

ocalypse

-0.79

ems

-0.78

urgy

-0.74

ograp

-0.73

idem

-0.73

uador

-0.72

POSITIVE LOGITS

 Hills

1.23

 Crusher

0.94

 Boulevard

0.94

 Peninsula

0.89

 Hilton

0.89

 Beverly

0.87

 Blvd

0.84

 Trees

0.83

 peninsula

0.80

 Bridge

0.78

Activations Density 0.027%

names of specific locations or cities

mentions of locations, particularly those in Beverly Hills and Pasadena

No Comments

No Known Activations