INDEX

Explanations

word sequences related to different locations, possibly names of cities or countries

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names of places and people

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

 Gors

-0.69

cause

-0.67

utherford

-0.65

oldown

-0.64

eatures

-0.63

ylene

-0.61

wat

-0.60

@#&

-0.59

votes

-0.59

yond

-0.57

POSITIVE LOGITS

Ã¨re

0.96

 Parish

0.84

arson

0.80

Ø©

0.78

igne

0.75

 Dame

0.74

ese

0.73

 Resort

0.70

ensis

0.70

idad

0.69

Activations Density 0.269%

word sequences related to different locations, possibly names of cities or countries

proper nouns, particularly names of places and people

No Comments

No Known Activations