INDEX

Explanations

words related to geographical locations or places

oai_token-act-pair · gpt-3.5-turbo

specific proper nouns, particularly names of locations and people

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

[+

-0.78

ONSORED

-0.75

anmar

-0.70

ð

-0.70

PET

-0.70

places

-0.68

values

-0.67

SHIP

-0.66

MIT

-0.63

TPPStreamerBot

-0.63

POSITIVE LOGITS

Et

1.18

iquette

1.01

ihad

0.99

obic

0.94

udes

0.87

alia

0.85

ues

0.84

roit

0.82

erno

0.81

aida

0.75

Activations Density 0.007%

words related to geographical locations or places

specific proper nouns, particularly names of locations and people

No Comments

No Known Activations