INDEX

Explanations

the location "New York" or variations of it

oai_token-act-pair · gpt-3.5-turbo

mentions of New York and related geographic references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

vol

-0.74

 Russ

-0.69

 Sear

-0.69

 Constantin

-0.69

conserv

-0.69

 Homo

-0.65

 Soph

-0.62

Geo

-0.61

ãĤ¦

-0.61

book

-0.60

POSITIVE LOGITS

 CITY

1.08

 YORK

1.07

FORE

0.92

ANE

0.90

ARD

0.82

INESS

0.82

ITED

0.81

amaru

0.79

STON

0.78

AMI

0.78

Activations Density 0.005%

the location "New York" or variations of it

mentions of New York and related geographic references

No Comments

No Known Activations