INDEX

Explanations

dates and locations mentioned in news articles

oai_token-act-pair · gpt-3.5-turbo

date and location references specific to news events

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

nesia

-0.61

wo

-0.59

 Hearts

-0.58

 Romans

-0.58

 typo

-0.57

 tune

-0.57

 treaties

-0.56

poke

-0.56

 Italians

-0.56

 unanswered

-0.51

POSITIVE LOGITS

ember

0.71

arity

0.71

tnc

0.69

 REUTERS

0.69

Aug

0.68

interstitial

0.68

window

0.66

thumbnails

0.65

Feb

0.63

EMBER

0.62

Activations Density 0.078%

dates and locations mentioned in news articles

date and location references specific to news events

No Comments

No Known Activations