INDEX

Explanations

words related to strong negative emotions such as loathing and hatred

oai_token-act-pair · gpt-3.5-turbo

words related to atheism or non-belief

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

Dek

-0.72

 Latvia

-0.70

TBA

-0.68

 UNHCR

-0.68

MO

-0.67

 Congo

-0.67

Zub

-0.66

Io

-0.66

 Lakes

-0.66

 Migration

-0.66

POSITIVE LOGITS

adoes

1.05

rette

1.01

reditary

0.89

adesh

0.88

istical

0.85

jriwal

0.84

hedral

0.83

artisan

0.83

ħĭ

0.80

ados

0.80

Activations Density 0.030%

words related to strong negative emotions such as loathing and hatred

words related to atheism or non-belief

No Comments

No Known Activations