INDEX

Explanations

significant events or people from different time periods

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

hyde

-0.69

ATURES

-0.69

vet

-0.69

INO

-0.68

plets

-0.65

ARA

-0.59

cil

-0.59

LY

-0.58

 Explain

-0.58

Features

-0.57

POSITIVE LOGITS

 century

1.23

day

1.20

 bunch

1.19

 millennium

1.17

decade

1.15

era

1.13

 year

1.13

 week

1.10

 night

1.08

 month

1.04

Activations Density 0.114%

significant events or people from different time periods

No Comments

No Known Activations