INDEX

Explanations

proper nouns, especially names of places or people, with a strong activation for the name "Lund"

oai_token-act-pair · gpt-3.5-turbo

mentions of specific names or places

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

qua

-0.67

 Akron

-0.66

chnology

-0.62

addock

-0.61

 conscious

-0.60

ASED

-0.59

ï

-0.59

ilic

-0.59

addon

-0.58

é¾įå¥ĳå£«

-0.58

POSITIVE LOGITS

gren

1.17

 Lund

1.13

sson

1.12

borg

1.04

qv

1.01

quist

0.92

holm

0.86

berg

0.85

lund

0.85

ieth

0.84

Activations Density 0.006%

proper nouns, especially names of places or people, with a strong activation for the name "Lund"

mentions of specific names or places

No Comments

No Known Activations