INDEX

Explanations

phrases involving expressing strong emotions or judgments towards others

oai_token-act-pair · gpt-3.5-turbo

second-person pronouns emphasizing direct address or accusations toward the reader

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

icy

-0.70

ice

-0.67

¿½

-0.66

ftime

-0.65

assembly

-0.64

 airs

-0.64

FEC

-0.60

UTC

-0.60

acular

-0.59

 Above

-0.58

POSITIVE LOGITS

're

1.57

've

1.34

'll

1.31

'd

1.16

tub

1.13

 guys

1.08

 know

0.92

owe

0.91

 guessed

0.89

RS

0.89

Activations Density 0.240%

phrases involving expressing strong emotions or judgments towards others

second-person pronouns emphasizing direct address or accusations toward the reader

No Comments

No Known Activations