INDEX

Explanations

web inputs prompting user engagement by using the word "Enter"

oai_token-act-pair · gpt-3.5-turbo

the word "Enter" 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 @ 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

 conviction

-0.72

 Archangel

-0.72

 laundering

-0.65

 backdrop

-0.65

 relayed

-0.62

 bombed

-0.60

 disposed

-0.59

packing

-0.59

 blow

-0.59

 gunned

-0.58

POSITIVE LOGITS

prise

1.74

prises

1.69

tainment

1.53

prising

1.46

tain

1.31

taining

1.28

prisingly

0.93

tained

0.89

TAIN

0.86

ymes

0.83

Activations Density 0.011%

web inputs prompting user engagement by using the word "Enter"

the word "Enter" in various contexts

No Comments

No Known Activations