INDEX

Explanations

references to historical periods or epochs

oai_token-act-pair · gpt-3.5-turbo

references to different historical eras

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

acts

-0.74

lations

-0.72

nil

-0.71

 drawer

-0.71

ãģį

-0.68

acted

-0.67

matter

-0.65

achu

-0.65

ergus

-0.64

yrights

-0.63

POSITIVE LOGITS

icals

1.05

icity

0.90

era

0.88

 epoch

0.84

Era

0.81

UTC

0.80

 nostalg

0.79

osaurs

0.75

 onward

0.71

 reckoning

0.71

Activations Density 0.028%

references to historical periods or epochs

references to different historical eras

No Comments

No Known Activations