INDEX

Explanations

references to time frames in terms of multiple years or centuries

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 ANGEL

-0.67

room

-0.64

iqu

-0.62

Redd

-0.59

WN

-0.58

ooter

-0.58

ebin

-0.57

Plan

-0.56

DV

-0.56

mber

-0.55

POSITIVE LOGITS

ago

0.96

oths

0.80

 ILCS

0.79

arily

0.76

worth

0.75

BCE

0.75

 millennia

0.74

olithic

0.73

arios

0.73

undo

0.71

Activations Density 0.055%

references to time frames in terms of multiple years or centuries

No Comments

No Known Activations