INDEX

Explanations

phrases related to specific locations, particularly those with the word "Grand" followed by numbers

oai_token-act-pair · gpt-3.5-turbo

references to the word "Grand" in various 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

Downloadha

-0.84

lihood

-0.81

tics

-0.74

ylon

-0.70

ople

-0.69

phrine

-0.68

PsyNetMessage

-0.68

ologies

-0.67

bin

-0.67

ijk

-0.67

POSITIVE LOGITS

father

1.24

mother

1.20

 Prix

1.19

parents

1.12

 Rapids

1.04

 Theft

1.00

iosity

1.00

master

0.98

pa

0.95

parent

0.93

Activations Density 0.012%

phrases related to specific locations, particularly those with the word "Grand" followed by numbers

references to the word "Grand" in various contexts

No Comments

No Known Activations