INDEX

Explanations

words related to locations, specifically those with 'oa' in them

oai_token-act-pair · gpt-3.5-turbo

specific geographical names and locations

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Chatt

-0.66

 quint

-0.64

 Vest

-0.63

 viol

-0.63

dro

-0.62

 bottom

-0.62

 rest

-0.61

adv

-0.60

 magnets

-0.60

SV

-0.60

POSITIVE LOGITS

oan

3.53

oa

3.31

OA

2.62

oat

1.99

aea

1.33

onut

1.30

oi

1.24

oha

1.21

oj

1.08

antha

1.02

Activations Density 0.015%

words related to locations, specifically those with 'oa' in them

specific geographical names and locations

No Comments

No Known Activations