INDEX

Explanations

terms related to geographical locations, specifically focusing on the word "Strait"

oai_token-act-pair · gpt-3.5-turbo

mentions of the Strait

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

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

 Newsp

-0.63

oric

-0.62

ale

-0.60

 Authorities

-0.59

mal

-0.58

 ballots

-0.57

 Otto

-0.55

 Lemon

-0.55

oration

-0.55

 retri

-0.55

POSITIVE LOGITS

pped

1.32

pping

1.26

ights

1.14

fing

1.11

pless

1.08

ÃŁ

1.00

ppy

0.99

ps

0.99

ppers

0.98

pper

0.97

Activations Density 0.034%

terms related to geographical locations, specifically focusing on the word "Strait"

mentions of the Strait

No Comments

No Known Activations