INDEX

Explanations

words related to the concept of "self" or possibly personal names

oai_token-act-pair · gpt-3.5-turbo

references to "elf" and related variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Cit

-0.75

 Alexis

-0.67

 reservation

-0.67

 traction

-0.67

Nap

-0.66

 reserve

-0.65

 Sand

-0.64

 Moder

-0.64

 drafts

-0.63

 subscribed

-0.63

POSITIVE LOGITS

elf

5.14

ELF

2.66

elf

1.37

Elf

1.30

gem

1.25

orc

1.13

alf

1.11

ikh

1.08

 elves

1.07

ela

1.06

Activations Density 0.039%

words related to the concept of "self" or possibly personal names

references to "elf" and related variations

No Comments

No Known Activations