INDEX

Explanations

dates represented in the format of years

oai_token-act-pair · gpt-3.5-turbo

significant dates and years in historical contexts

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

 Skinner

-0.66

bip

-0.64

 monop

-0.64

acci

-0.62

ãĤ¤ãĥĪ

-0.62

oi

-0.61

warts

-0.58

 peaked

-0.57

rouse

-0.57

urden

-0.57

POSITIVE LOGITS

GV

0.72

rd

0.71

é¾į

0.71

å¹

0.70

âĢ¢âĢ¢

0.66

roxy

0.66

EW

0.65

GAN

0.65

 Nicol

0.65

UTERS

0.64

Activations Density 0.199%

dates represented in the format of years

significant dates and years in historical contexts

No Comments

No Known Activations