INDEX

Explanations

movie titles enclosed in quotation marks

oai_token-act-pair · gpt-3.5-turbo

quotations from various sources

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

²¾

-0.79

senal

-0.75

¬¼

-0.73

ĻĤ

-0.72

ife

-0.69

cha

-0.67

comings

-0.66

acas

-0.65

isl

-0.64

nex

-0.63

POSITIVE LOGITS

/"

0.85

 referring

0.82

aka

0.79

 meaning

0.78

SPONSORED

0.78

 implying

0.74

 namely

0.72

 according

0.71

 referencing

0.71

agy

0.70

Activations Density 0.061%

movie titles enclosed in quotation marks

quotations from various sources

No Comments

No Known Activations