INDEX

Explanations

phrases related to a person or object entering a particular situation or agreement

oai_token-act-pair · gpt-3.5-turbo

instances of the word "entered" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

axter

-0.84

 Tycoon

-0.75

ggies

-0.68

killer

-0.66

 opio

-0.65

wig

-0.64

 sled

-0.63

pedia

-0.62

shoot

-0.62

visor

-0.61

POSITIVE LOGITS

prises

0.96

 entered

0.92

prise

0.89

 Entry

0.81

prising

0.75

 entering

0.74

 ENTER

0.73

 INTO

0.72

 enter

0.71

erville

0.69

Activations Density 0.011%

phrases related to a person or object entering a particular situation or agreement

instances of the word "entered" in various contexts

No Comments

No Known Activations